Storage Unpacked Podcast

View Details

In this episode, Chris talks to Sharad Kumar, Field CTO at Qlik about the value of good-quality data when developing AI solutions. Much of the current discussion around AI and large-language models (LLMs) is focused on the infrastructure and the significant expense needed to build and train generative AI. However, as the revelation of DeepSeek shows, the industry trend will see models commoditise and become cheaper to train and run.

If infrastructure and software become quickly affordable, what is the differentiator for businesses? The answer is clearly their data. Data has value to an enterprise, but only if it is in an acceptable format. That means being of high quality and in terms of how Qlik operates, a trusted resource.

During the conversation, Sharad explains the six metrics of the Talend Trust Score, a methodology that measures the value of data based on Diversity, Timeliness, Accuracy, Security, Discoverability and Consumability. He explains how the Trust Score is calculated, but more importantly, how businesses can build a framework to continually improve the quality and value of their data resources.

More information on Qlik can be found on the company website – here. Sharad mentions the user conference taking place in May, details of which can be found here. Finally, Sharad references the Qlik LinkedIn page, which can be found here.

Elapsed Time: 00:47:47

Timeline* 00:00:00 – Introductions * 00:01:46 – Data is the value piece within AI, not infrastructure * 00:02:27 – What is occurring within the AI market? * 00:04:25 – The future will be a mix of AI model types and sizes * 00:05:20 – Will businesses build or buy models? * 00:07:10 – How will agentic AI architectures work? * 00:10:30 – Customers need to focus on data quality * 00:12:44 – Both training and RAG data needs to be high quality * 00:14:40 – Agentic AI wil be intent-driven * 00:16:43 – What does good data look like within an enterprise? * 00:19:28 – Qlik has a 6-dimensional trust score * 00:26:11 – How do customers calculate their trust score? * 00:30:09 – Is AI driving better data quality? * 00:34:51 – Qlik can help customers develop a data improvement programme * 00:37:36 – Qlik brings “product thinking” to data * 00:38:56 – Where are businesses on the AI journey? * 00:41:12 – How is improving data quality driving improving AI benefits? * 00:42:26 – AI could be applied to fix data quality problems

View Details

In this episode, Chris discusses the options available to storage system vendors when building modern storage appliances, with Bill Basinas, Senior Director, Product Marketing at Infinidat. The conversation derives from an observation on architectural choices, following the move to AMD processors from Intel for the latest G4 systems built by Infinidat. AMD offers a greater core count per processor compared to Intel, allowing Infinidat to move to single socket designs, while gaining improvements from PCIe 5.0 and DDR5 memory.

Ultimately, this discussion highlights how modern storage system design can take standardised components and build flexible architectures, implementing most features in software. For Infinidat, that could mean expanding its range of solutions for smaller enterprise requirements, or building out products specifically for Edge use cases.

Although Bill did not reveal any future plans, the implication is clear – watch this space for future evolution of the InfiniBox architecture to a wider and more varied set of hardwaree configurations.

Elapsed Time: 00:37:13

Timeline* 00:00:00 – Intros * 00:01:15 – How do vendors choose the hardware components for storage systems? * 00:02:30 – What are the main (storage) technology challenges for customers? * 00:04:08 – Customers want predictable data features * 00:05:55 – Capacity demand continues to grow relentlessly * 00:07:30 – Infinidat features are built into software * 00:09:35 – Most AI requirements wil run on existing performance storage * 00:11:20 – Modern hardware provides significant flexibility for system design * 00:15:00 – AMD gives access to single and high core-count processors * 00:16:10 – PCIe 5.0 provides for faster SSDs and power efficiency * 00:18:46 – Infinidat has introduced smaller form-factor solutions * 00:21:32 – Multiple cores will always get used! * 00:25:53 – Infinidat G4 architecture provides for in-place controller upgrades * 00:28:22 – Storage arrays should become more “virtual” * 00:34:10 – Data services implementations are very different between vendors * 00:35:55 – Hybrid architecture still has value in the Infinidat world * 00:36:20 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 258 – Introducing Infinidat G4, InfuzeOS 8 and InfiniSafe ACP * #202 – Enterprise Storage Consolidation with Phil Bullinger from Infinidat * Infinidat adds customer value with SSA Express and improved SSA capacity

View Details

In this episode, Chris discusses the enduring benefits of centralised storage, particularly with reference to storage virtualisation, with Dan Kogan, VP of Enterprise Growth and Solutions and Cody Hosterman, Senior Director of Product Management, both from Pure Storage.

Centralised or shared storage has been around for over 30 years, providing efficiencies in infrastructure and operational management. In the virtualisation context, centralisation provides the ability to abstract workloads from the hypervisor and add flexibility and data management features to a centrally managed platform. Vendors, such as Pure Storage, have invested resources in making centralised storage efficient, while also providing significant security benefits that couldn’t be achieved with an HCI model.

Although this discussion was intended to focus on centralisation, the ultimate conclusion of the conversation is to realise that centralised storage is a precursor to storage-as-a-service. This is where the industry is headed, whether using on-premises or public cloud infrastructure.

Elapsed Time: 00:35:34

Timeline* 00:00:00 – Intros * 00:01:17 – Shared or Centralised Storage has become a perpetual feature of the data centre * 00:02:00 – Where did centralised storage come from? * 00:03:03 – VMware introduced compute efficiencies, centralised storage does the same * 00:05:20 – Centralised storage now incorporates block, file and object protocols * 00:07:10 – HCI was probably the biggest “challenge” to centralised storage * 00:13:04 – Centralisation is bringing additional consolidation benefits * 00:15:55 – Centralisation provides significant operational benefits * 00:17:36 – Integrated storage (HCI) is inherently insecure compared to centralised storage * 00:22:31 – Data mobility is a key requirement of modern enterprises * 00:29:11 – Centralised storage is driving us towards storage-as-a-service. * 00:31:10 – Storage is becoming an “endpoint” * 00:32:31 – Wrap Up

Related Podcasts & Blogs* Analysis: Storage vendors assist in the optimisation of VMware workloads

View Details

In this episode, Chris is in conversation with Jeb Horton, SVP Global Services at Hitachi Vantara, discussing the capabilities of Hitachi Vantara’s Global Services offerings, which deliver infrastructure management and infrastructure as a service to its customers.

In addition to EverFlex, Hitachi Vantara has a long history of managed services capabilities that span more than just outsourced storage. As Jeb explains, the company also manages storage infrastructure from other vendors, in addition to non-storage systems.

The interesting aspect of this discussion is the complex nature of the interaction between customers and Hitachi. Solutions offerings aren’t merely “transactional”, but have a human aspect and are tailored to meeting the specific goals of the customer. This conversation explores some of the nuances of working with customers to transfer the burden of infrastructure management to Hitachi, enabling businesses to focus on more strategic opportunities.

To learn more about Hitachi Vantara check out the Infrastructure as a Service section on the Hitachi website – https://www.hitachivantara.com/en-us/services/infrastructure-as-a-service.

Elapsed Time: 00:48:02

Timeline* 00:00:00 – Intros * 00:01:43 – What is “Infrastructure as a Service”? * 00:03:25 – What else to customers want from a service (other than cost saving)? * 00:05:20 – Public cloud has increased the appetite for service-based consumption * 00:06:24 – What is the core of the Hitachi Vantara services offering? * 00:07:14 – Hitachi added automation into a “services platform” * 00:10:26 – The human aspect involves skills but also relationships * 00:12:20 – A service contract involves a detailed commercial model * 00:13:51 – Service also means service levels and agreements * 00:16:53 – Cloud is transactional, what is Hitachi’s “value add”? * 00:19:45 – Data has value, which is the focus of service offerings * 00:22:26 – How does Hitachi help government institutions? * 00:26:50 – What sort of data issues does Hitachi deal with? * 00:28:33 – Data and AI will be a key issue to manage * 00:30:40 – How does the engagement process work with Hitachi (and what is EverFlex)? * 00:37:15 – What are real-world examples of Hitachi customers and requirements? * 00:46:51 – Wrap Up

Related Podcasts & Blogs* Hitachi Vantara Microsite * Storage Unpacked 260 – Hitachi VSP One Updates with Dan McConnell * Storage Unpacked 254 – Announcing VSP One and Hitachi Vantara Reorganisation

View Details

In this recording, Chris talks to Subbiah Sundaram, SVP of Products at HYCU, Inc. about the 2024 edition of the HYCU State of SaaS Data Resilience Report. The report surveys customers to understand the gaps in perceived and actual data protection for SaaS platforms and the results are quite surprising. Subbiah walks through the top four findings, covering the understanding of the pervasive nature of SaaS in modern business, perceptions of data protection and the unexpected risks created by SaaS platforms.

HYCU provides a robust and comprehensive approach to SaaS data protection, called R-Graph, part of R-Cloud. We’ve covered these products in previous podcasts, shown in the related content section below. We recommend downloading the report, which can be found here – The State of SaaS Data Resilience in 2024. Details on R-Graph can be found here – R-Graph.

Elapsed Time: 00:39:59

Timeline* 00:00:00 – Intros * 00:01:39 – What is the SaaS Resiliency Report for 2024? * 00:02:23 – There are over 35,000 global SaaS applications * 00:04:11 – SaaS has become embedded in business process * 00:05:02 – Businesses underestimate SaaS applications by 10x * 00:06:29 – Businesses don’t realise SaaS data isn’t protected like on-premises * 00:09:20 – 61% of data breaches occur through SaaS platforms * 00:13:40 – Businesses assume cloud platforms protect their data * 00:15:18 – The reasons for data restoration are multi-fold and business related * 00:17:47 – 75% of critical infrastructure (identity management) was not being protected * 00:19:21 – All credentials management systems operate slightly differently * 00:22:30 – Business process creates historical security exceptions * 00:26:09 – use R-Graph to discover your application dependencies * 00:27:42 – Protect your identity management systems * 00:31:09 – R-Cloud enables anyone to add data integrations for backup * 00:32:26 – Protect your endpoints, protect your data, protect your customer data * 00:34:18 – Where does SaaS data protection go next? Tracking behaviour * 00:37:14 – R-Cloud can be used for cross-environment data seeding * 00:39:12 – Wrap Up

Related Podcasts & Blogs* Data Unpacked 006 – Introducing HYCU R-Cloud * Data Unpacked 004 – Reflections on Data Management, Security & Protection With HYCU CEO Simon Taylor * Research Note: HYCU extends SaaS Integration with R-Scout and Generative AI * HYCU expands SaaS and IaaS backup with protection for AWS Infrastructure as Code * HYCU tackles the SaaS data protection challenge with the announcement of R-Cloud

View Details

In this podcast episode, Chris is in conversation with Jeffries Briginshaw (Head of EMEA Government Relations at NetApp) and Adam Gale (CTO for AI & Cyber Security, NetApp) discussing the EU AI Act and the regulation of artificial intelligence across the world. The EU AI Act is an early introduction into the regulation of the use of AI by businesses within their engagements and interactions with customers. As explained in this conversation, there are classifications of AI types and within that, restrictions on what businesses are permitted to implement based on those categorisations. Some AI usage will be banned, while others will require human intervention and close monitoring.

How should your business engage with AI and ensure compliance with the act? Listen to the discussion for more details. As mentioned in the recording, for details on what NetApp can offer, point your favourite browser to https://www.netapp.com/artificial-intelligence/ to learn more.

Elapsed Time: 00:52:17

Timeline* 00:00:00 – Intros * 00:01:19 – Why should we be regulating AI? * 00:02:30 – What will the impacts of AI be on personal and work life? * 00:03:55 – What if we get regulation wrong? * 00:05:30 – What happens if AI goes wrong, such as data poisoning? * 00:09:04 – Existing EU/UK law has been successful at regulation (GDPR) * 00:10:25 – What is the EU AI Act? * 00:11:46 – “Prohibited Practices” will be banned from 2025 * 00:14:00 – How will the use of business in AI be regulated? * 00:18:05 – The EU AI Act appears to focus on protection for individuals * 00:20:56 – EU citizens are broadly positive to AI – if it is successfully regulated * 00:21:52 – Compliance has an overhead – in terms of hard costs (developers) * 00:25:20 – What are the penalties for not complying with the EU AI Act? * 00:29:50 – What about the rest of the world – the US and elsewhere? * 00:35:10 – Could we see “cross-border” complexity? * 00:37:40 – What are the technology implications for AI regulation? * 00:40:07 – Should businesses be demonstrating their AI compliance? * 00:44:03 – What does NetApp offer customers to help AI compliance? * 00:47:38 – AI will require a “big red stop button” * 00:50:00 – Wrap Up

View Details

In this podcast episode, Chris discusses the platform update announcements from Pure Accelerate 2024 with Prakash Darji, VP and GM of the Digital Experience BU at Pure Storage. The new features focus on usability and operational enhancements, including AI-based features and support for AI workloads. Highlighted in this discussion are:

Fusion automation enhancements for fleet management and individual arrays
New Generative AI Copilot for storage to provide querying capabilities and advice
Evergreen//One for AI – an AI-level tier of Storage-as-a-Service
NVIDIA SuperPOD Ethernet Certification
Secure application Workspaces using Portworx
Cyber Recovery and Resilience SLAs
Security Assessment SLA
AI-Powered Anomaly Detection enhancements
Site Rebalance SLA
AI-Powered Reserve Expansion recommendations

As the list shows, there are lots of new updates to make the management and operation of a Pure Storage fleet more efficient and easy. As Prakash explains the reasoning behind the features, it is clear that AI is being used to deliver simplicity, while the platform will provide support for customers wanting to build AI-focused workloads.

To learn more, follow the news from Pure Accelerate 2024 here (link). Prakash mentions two blog posts, which can be found here – Ransomware is a Darwinian Problem That Will Never Be Solved and Editorial: Why Centralised Storage Refuses to Go Away.

Elapsed Time: 00:38:33

Timeline* 00:00:00 – Intros * 00:00:51 – It’s not all about AI! * 00:01:34 – What changes have been announced to the Pure Storage platform? * 00:02:37 – New features include cybersecurity enhancements and simplicity of management * 00:03:30 – How do we manage systems at scale? * 00:04:27 – Applications need policy management * 00:05:08 – Fusion has been enhanced to enable array or fleet management at the same time * 00:08:10 – Pure is introducing a GenAI Copilot in preview * 00:12:19 – Evergreen now has an AI storage-as-a-service tier * 00:14:00 – Pay for performance and capacity is a feature of Evergreen * 00:15:55 – SuperPod certification for Ethernet is coming to Pure Storage arrays * 00:16:40 – There must be many Jensen clones * 00:18:27 – Pure is introducing secure application workspaces using Portworx * 00:22:32 – New cybersecurity features include a security assessment for configuration settings * 00:23:19 – There is also a security SLA for fixing and certificating security settings * 00:24:01 – The AI Copilot will also recommend security improvements * 00:24:32 – Anomaly detection is now performance-based, looking at typical profiles * 00:30:45 – Reserve expansion recommendation is now AI-powered * 00:31:55 – Reserve commit across sites can now be rebalanced once per year * 00:33:40 – It’s easy for storage to become fragmented between sites * 00:36:27 – When will the new features be made available? * 00:37:45 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 259 – Sustainable Storage in the World of AI with Shawn Rosemarin * Storage Unpacked 257 – The Future of Data Storage in the Enterprise * Storage Unpacked 252 – A Vision of Storage Future with Coz from Pure Storage * Storage Unpacked 251 – Modernising Storage as a Service with Prakash Darji * Pure Storage Microsite * X-Ray: Pure Storage, Inc.

View Details

In this podcast episode, Chris catches up with Dan McConnell, Senior VP for Product Management at Hitachi Vantara. The company recently announced VSP One Block, a new mid-range appliance for block storage. This follows on from two product announcements in April, which we covered in this Research Note, and the restructuring of Hitachi Vantara announced towards the end of last year (see this Research Note).

Dan discusses VSP One Block, an appliance that targets mid-range storage requirements. He also covers VSP One SDS, a software-defined solution which runs in AWS and on-premises. The third product announcement covers file, with VSP One File, the latest iteration of the technology that came from the BlueArc acquisition over a decade ago.

You can find out more about the Block Storage Appliance here (link and here). Details on the VSP One SDS announcement can be found here (link), which includes details on VSP One File.

Elapsed Time: 00:15:29

Timeline* 00:00:00 – Intros * 00:01:18 – April 2024 announcement – VSP One SDS & VSP File * 00:02:00 – Hitachi blog products use SVOS * 00:03:13 – VSP one SDS is scale-out * 00:03:51 – VSP File is the evolution of previous file-based products * 00:05:16 – The VSP One family introduces consistent management & hybrid support * 00:06:24 – EverFlex introduces multiple consumption models * 00:08:30 – VSP Block 20 is the next generation mid-range storage array * 00:10:10 – Dynamic Carbon Reduction optimises power usage by workload demand * 00:12:03 – What comes next? * 00:13:15 – Cloud storage products shouldn’t be a “lift and shift” * 00:14:47 – Wrap Up

Related Podcasts & Blogs* Research Note: Hitachi Vantara VSP One reaches GA * Research Note: Hitachi Vantara reorganises and announces VSP One Platform * Hitachi Vantara Microsite * X-Ray: Hitachi Vantara * Storage Unpacked 254 – Announcing VSP One and Hitachi Vantara Reorganisation with Gary Lyng

View Details

In this episode, Chris discusses the topic of building sustainable storage solutions with Shawn Rosemarin, Global VP of Customer Engineering at Pure Storage. AI and specifically Generative AI (GenAI) has become a hot topic over the past 12 months. Businesses are looking at projects to use AI internally for productivity gains, but also to drive additional business.

However, AI is still relatively expensive and requires huge volumes of training data. Training is an ongoing process that must react to changes in the data landscape, such as rights and permissions, and government regulation. With AI hardware being so expensive, it’s important to get the storage piece right, and that means having a scalable and cost effective solution. Shawn details how Pure Storage has focused on two aspects. First, the hardware, where DFMs (direct flash modules) have reached 75TB, with commitments to deliver 150TB and 300TB drives in the next few years. Second, the software management capability delivered through Purity, the operating system of Pure Storage hardware.

It’s clear that building cost and power-efficient flash devices will be a challenge for the wider industry, where the focus lies with consumer devices. Pure Storage believes it is well positioned to help customers and potentially hyper-scalers in their goals to deliver efficient storage for AI.

As Shawn highlights, this topic and more will be discussed at Pure Accelerate, to be held in Las Vegas from 18-21 June 2024. Check out the website where you can learn more.

Elapsed Time: 00:52:08

Timeline* 00:00:00 – Intros * 00:01:44 – We’ve been quiet on the topic of AI * 00:03:10 – AI has become cost-effective (sort of) * 00:04:00 – Efficient AI is a 10-15 year journey * 00:05:22 – AI technology needs to be efficient due to the resource demands * 00:06:41 – Data growth is currently growing at 30% per annum * 00:07:31 – Early mover may not be the best move with AI * 00:08:16 – 149 foundational models were released in 2023 * 00:09:10 – Businesses will want to merge public and private data * 00:10:40 – Results accuracy is super-important * 00:13:30 – Trusted AI will be adopted in areas like security & vehicle evasive manoeuvres * 00:15:10 – Where will AI models be developed? * 00:16:37 – Model retraining will be required due to changing data ownership & permissions * 00:18:30 – Model training also needs to be resource efficient * 00:19:49 – $100 million to do the basic training of an AI model * 00:22:26 – How do you feed GPUs with adequate data to run at 100% * 00:24:10 – Edge devices could be used for AI processing * 00:25:18 – How will data centres need to evolve for AI? * 00:28:08 – Sustainability, regulation and jobs will all be issues in AI deployment * 00:31:05 – With HPC, many users built bespoke systems and that’s a problem for AI * 00:33:45 – How will businesses “industrialise” their AI projects? * 00:36:46 – Storage density will help resolve the operational issues of AI storage * 00:38:19 – SSD vendors’ main market is 2TB consumer SSDs * 00:39:32 – 300TB drives are great, but how will software manage the hardware? * 00:41:51 – Pure Storage DFMs will grow exponentially in capacity * 00:42:43 – Hardware engineering is cool again! * 00:45:15 – How will the hyper-scalers deal with massive storage growth? * 00:51:30 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 257 – The Future of Data Storage in the Enterprise * Storage Unpacked 252 – A Vision of Storage Future with Coz from Pure Storage * Storage Unpacked 245 – Design Strategies for 300TB Flash Drives with Shawn Rosemarin from Pure Storage * Dude, Here’s Your 300TB Flash Drive!

View Details

In this episode, Chris talks to Infinidat CMO, Eric Herzog. Infinidat has announced one of the biggest upgrades in eight years, with the release of InfiniBox and InfiniBox SSA G4, the fourth generation of enterprise-class storage. Accompanying the new hardware is an upgrade to InfuzeOS, the Infinidat storage operating system, and a new feature for InfiniSafe – Automated Cyber Protection, or ACP.

Infinidat has upgraded both the InfiniBox and InfiniBox SSA platforms with a generation 4 release that includes a switch to AMD processors. Using the EPYC 9554P enables Infinidat to use a single-socket design, while gaining from the move to DDR5 system memory and PCIe 5.0 I/O. The savings to the customer are space, power and cooling. The AMD move also enables Infinidat to release new hardware configurations, including a 14U rack-mount solution for edge data centres, rather than just the custom rack used to ship existing products.

InfuzeOS gains an upgrade to version 8, with support for InfuzeOS in the public cloud on Microsoft Azure (AWS was announced last year). The new hardware and software improvements result in a 2x performance gain for customers. One final announcement covers InfiniSafe and the ability to automate snapshots through the integration of cyber-detection technology with InfiniBox and InfiniBox SSA. Customers can now automate the creation of immutable snapshots if their SIEM or SOAR platform detects malicious activity. This capability reduces the size of the threat window and the potential volume of data needing recovery, should a breach occur.

There’s a lot more detail in the podcast, so go ahead and listen! For more information on any of the announcements in this podcast episode, visit https://www.infinidat.com/.

Elapsed Time: 00:47:27

Timeline* 00:00:00 – Intros * 00:01:00 – What’s new with Infinidat? G4 hardware, InfuzeOS updates and InfiniSafe ACP * 00:01:40 – G4 -new platform, both hybrid & all-flash, using AMD processors * 00:02:50 – Processor choice is available, but software is the key * 00:04:00 – InfuzeOS 8.0 is compatible with previous hardware generations * 00:04:40 – How has the physical specification of systems changed? * 00:07:15 – 14U option now available for use in standard racks * 00:10:00 – Why new form factor? Increased TAM * 00:11:00 – New controller upgrade programme introduced – Mobius * 00:12:15 – In-place upgrades are more practical with flash systems * 00:15:29 – What is InfiniVerse? * 00:18:25 – Fleet Management is now table stakes – and a differentiator * 00:20:17 – InfuzeOS is now available in AWS and Azure * 00:24:45 – Why use a cloud SDS solution – portability * 00:27:09 – InfiniSafe – what is Automated Cyber Protection? * 00:29:02 – Guaranteed immutable snapshots & recovery times * 00:33:39 – Dynamic snapshots based on threat identification reduces threat windows * 00:37:00 – ACP provides an holistic approach to data security * 00:40:41 – InfiniSafe cyber-protection now scans VMware virtual machine datastores * 00:44:00 – What is the availability of all the new offerings? * 00:45:30 – Live demos and Webinars are coming over the next few months * 00:46:42 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 250 – Infinidat announces SSA Express and higher capacity SSA II * Storage Unpacked 247 – Infinidat announces InfuzeOS Cloud Edition and InfiniSafe Cyber Detection * #231 – Introducing Infinidat InfiniBox SSA II * #227 – Infinidat InfiniGuard Enhancements with Eric Herzog * Infinidat Microsite * Infinidat adds customer value with SSA Express and improved SSA capacity * The Quiet Success of Infinidat

View Details

In this sponsored episode, Chris talks to Fred Lherault and Larry Touchette from Pure Storage on the evolution of storage in the enterprise and the impacts on storage administration. The conversation is divided into three areas focusing on the customer, the administrator and the business.

From the customer’s perspective, the requirements of on-premises data centre storage have changed significantly. Users expect resources to be deployed on demand, using APIs, CLIs or a GUI, without the intervention of a storage administrator. The self-service aspect is also aligned with 100% availability, an expectation that has evolved from the public cloud. End users have less interest in the hardware itself, but instead focus on metrics (IOPS, latency, throughput) and see storage as an endpoint to be consumed.

The role of the storage administrator has evolved to be one similar to that of a product manager. The administration role is much more focused on ensuring storage is available and operating efficiently, rather than on the mundane task of provisioning resources. This means keeping close control on capacity growth, upgrades and patching.

For the business, costs and efficient consumption models are key. With 30-40% annual growth in consumed terabytes, year-on-year costs need to decline, while systems must become more power, space and cooling efficient. Pure Storage has introduced Pure1 and Fusion, tools for the business and administrators to ensure that the storage infrastructure operates efficiently and meets the SLAs expected by internal customers.

During the discussion, we highlight Pure Storage’s annual user conference, Accelerate, which will take place in Las Vegas between June 18th and 21st. Here is a list of some useful related content that discusses the evolution of storage in the data centre.

Storage Unpacked 252 – A Vision of Storage Future with Coz from Pure Storage
Storage Unpacked 251 – Modernising Storage as a Service with Prakash Darji from Pure Storage
Storage Unpacked 248 – FlashArray R4 Announcements From Pure Accelerate 2023
Storage Unpacked 245 – Design Strategies for 300TB Flash Drives with Shawn Rosemarin from Pure Storage
The Great Cloud Repatriation Debate – Data Storage
Pure Storage announces Disaster Recovery-as-a-Service, SLA guarantees and operational cost rebates
Delivering True Storage-as-a-Service
Pure//Launch – Pure Storage Announces Fusion and Portworx Data Services

Elapsed time: 00:49:53

Timeline* 00:00:00 – Intros * 00:02:25 – How has storage management changed over the last two decades? * 00:03:07 – What are the modern storage requirements of enterprise customers? * 00:04:32 – The speed and agility of the public cloud is driving on-premises expectations * 00:06:40 – There is a mix of customer maturity in the enterprise * 00:09:48 – Customers expect less focus on hardware and more on metrics of delivery * 00:12:00 – Sustainability – including power costs – are increasingly important to customers * 00:13:05 – Automation – via GUI, API and CLI is expected, to reduce delivery times * 00:14:51 – Businesses expect 100% uptime, with no downtime requirement for upgrades * 00:17:01 – Storage “arrays” are now virtual, as data outlives the hardware * 00:18:46 – Do storage administrators now have an easier job? * 00:20:33 – Pure Storage takes some of the admin burden off the customer * 00:22:13 – Admins need to manage infrastructure, while providing access to the technology * 00:24:52 – How do businesses manage the financial demands of growing storage needs? * 00:27:27 – Modern consumption models are driven by architectural features * 00:29:16 – Pure Storage has operational processes to manage customer on-demand consumption * 00:33:15 – Efficient resource management is analogous to retail stock control * 00:34:00 – Pure1 and analytics tools provide the capability to efficiently model workload placement * 00:36:53 – Modern storage has many internal management functions that need AI/ML planning * 00:39:00 – So what should storage vendors be delivering, as minimum functional requirements? * 00:40:39 – Pure hardware and software is intrinsically linked * 00:41:45 – As flash improves, vendors like Pure can address many more performance & cost use cases * 00:45:00 – Pure systems started at 5.5TB, now into multi-petabytes * 00:48:10 – Wrap Up

View Details

In this episode, Chris chats to Rick Kutcipal, “At-Large Director” with the SCSI Trade Association. The topic of conversation is the adoption of SAS media (both HDDs and SSDs) by hyper-scale customers that include public cloud vendors and companies such as Meta. Market perception implies that NVMe-based drives are taking over the world, but that’s far from the truth. As Rick explains, some 90% of exabytes shipped on SSDs and HDDs are still using the SAS interface. SAS scales much better (in terms of drives in systems) than NVMe, while offering a competitive price point when looking at “slot cost”.

There’s a lot of detail to digest in this discussion. It touches on some novel features of HDDs, for example, including Depop and Command Duration Limits. What is clear from the conversation is the longevity of SAS into the future, even as the transition to flash-based media continues.

To learn more about the SCSI Trade Association, check out their website at https://www.snia.org/groups/sta-forum. You can also find them on Linkedin – here.

Elapsed Time: 00:33:09

Timeline* 00:00:00 – Intros * 00:01:30 – Hyper-scalers are big users of SAS devices * 00:02:55 – Refresher – What are SAS and SATA? * 00:04:55 – What are storage requirements for Hyper-scalers? * 00:05:40 – Requirements differ by area (engineers, operations and architects) * 00:07:15 – Small percentage savings make a big difference to Hyper-scalers * 00:08:45 – SAS scales to thousands of drives, with built-in management * 00:10:00 – Certain features have been added specifically for Hyper-scalers * 00:12:10 – I/O density continues to decline with HDD capacity increases * 00:13:55 – Drive systems can be a mix of NVMe and SAS/SATA drives * 00:15:30 – Reliability is critical, to avoid data centre interventions * 00:18:15 – Scale is only achievable with SAS * 00:19:35 – The supplanting of HDDs by SSDs is debatable * 00:21:30 – Large-scale SSDs are seeing the same issue as large HDDs * 00:23:30 – Tiering will continue to be important within the storage industry * 00:25:00 – Exabytes shipped still shows 90% remains behind SAS infrastructure * 00:27:50 – Power comparisons between SSD and HDD are not clear cut * 00:29:35 – Hyper-scalers focus on “slot cost” * 00:30:30 – What businesses are using SAS solutions – Meta * 00:31:55 – Wrap Up

Related Podcasts & Blogs* #238 – SAS 24GB+ Updates with Rick Kutcipal * #74 – All About Serial Attached SCSI with Rick Kutcipal

View Details

In this episode, Chris is in conversation with Ryan Farris (VP Product and Product Marketing) and Brandon Whitelaw (VP Cloud and Strategic Partnerships) at Qumulo. As the IT world becomes ever more focused on a hybrid cloud model, file storage becomes increasingly important, due to the legacy of applications and data already written to work with file servers.

However, file storage in the public cloud doesn’t have the same features and flexibility as native object or block storage solutions. Most solutions operate and feel like on-premises infrastructure, lifted and shifted to the public cloud. So, what should file storage look like in a hybrid cloud world? Listen to find out!

As we were recording this podcast, Qumulo was planning a big product announcement of the “Scale Anywhere” platform. New features include native support on Microsoft Azure and distributed file system capabilities. Check out the details here – https://qumulo.com/a-new-era-for-cloud-file-storage/

Elapsed Time: 00:39:29

Timeline* 00:00:00 – Intros * 00:01:22 – How are customers moving their around the enterprise today? * 00:03:11 – How is data workflow changing? * 00:05:52 – Is most new data unstructured? * 00:07:30 – Unstructured data has become mission critical content * 00:08:50 – Where is new data being created? * 00:11:58 – File is becoming more important for cloud apps, but are they all useful? * 00:14:55 – Re-writing to an object store is not a cost efficient plan * 00:16:49 – Enterprises could have hundreds of apps writing to a central NAS platform * 00:18:55 – Object storage users expect worse latency than file systems * 00:20:40 – What data mobility solutions and requirements are we going to see? * 00:24:10 – NAS for hybrid cloud needs a new set of requirements * 00:27:05 – Requirements, presentation layer, endpoint capability, data moving capability * 00:30:15 – How does Qumulo deliver to these requirements? * 00:35:37 – What has Qumulo just announced? * 00:38:45 – Wrap Up

Related Podcasts & Blogs* #189 – The Quiet Success of Software-Defined Storage * Data Mobility – Global/Scale-out Data Platforms

View Details

In this live episode, recorded at Hitachi Exchange in Paris, Chris chats to Gary Lyng, VP of Products and Solutions at Hitachi Vantara. The company recently announced the VSP One platform, plus some organisational changes that will take Hitachi Vantara back to a focus on core infrastructure. This recording dives into the strategy behind VSP One, before new products and services follow in 2024.

We covered the VSP One announcement and reorganisation details in a blog post, available here – https://www.architecting.it/blog/hitachi-vsp-one/

More information on Hitachi Vantara is available through our X-Ray eBook (subscription required).

Elapsed Time: 00:31:37

Timeline* 00:00:00 – Intros * 00:01:50 – The 100% availability guarantee is 20+ years old * 00:02:40 – What is the VSP One announcement? * 00:04:10 – With many silos, infrastructure has become complex * 00:07:05 – The current announcement is a strategy, products due in 2024 * 00:08:10 – Modern storage requirements have evolved * 00:09:55 – Reliability, consistency and availability are key attributes of modern systems * 00:13:30 – GenAI and analytics are driving data volumes * 00:14:55 – Can data be culled or at least tidied? * 00:15:50 – Humans like to keep “stuff” * 00:19:45 – Modern IT systems can never be offline * 00:21:20 – Cloud now has high performance instances that can build virtual SANs * 00:24:20 – DLM is back, but in a new way, as data value can increase over time * 00:25:05 – Keeping data “forever” doesn’t really mean forever * 00:27:00 – Hitachi Vantara has restructured to move capabilities to Hitachi Digital Services * 00:30:50 – Wrap Up

Related Podcasts & Blogs* Hitachi Vantara reorganises and announces the VSP One platform * #160 – Updates on Hitachi Ops Center with Stan Stevens * #154 – Hitachi Vantara VSP E990 NVMe Midrange Appliance

View Details

In this episode, Chris chats with Abel Gordon, Chief System Architect at Lightbits Labs, discussing the challenges and benefits of building a virtual storage area network (SAN) on public cloud infrastructure. Lightbits originally developed the NVMe/TCP protocol and uses this feature to build virtual SANs using public cloud instances. This is a topic we first looked at in episode #210, so it’s good to get a practitioner’s experience.

Modern public cloud now features fast networking, low-latency NVMe and high-performance virtual and physical instances. Unfortunately, NVMe devices are ephemeral and any provisioned storage in the cloud is charged at full capacity. For users of on-premises SANs, the lack of thin provisioning may be an unwelcome surprise.

Why build a virtual SAN, other than to save storage costs? There’s a lot more involved, including delivering resiliency, scalability, targeted performance and capacity. Abel discusses the benefits, then goes on to enumerate the challenges involved when building on vendor-owned infrastructure. Finally, the discussion moves on to how Lightbits’ software is deployed and operated, including the managed application capability in Microsoft Azure.

For more information on Lightbits Labs, visit the company website at https://www.lightbitslabs.com/

As Abel, suggests you can contact him on LinkedIn or email him directly at abel@lightbitslabs.com.

Elapsed Time: 00:51:15

Timeline* 00:00:00 – Intros * 00:02:07 – Why build a virtual SAN in the public cloud * 00:04:30 – SANs balance out and fully exploit available performance resources * 00:06:36 – Public cloud charges for performance and capacity * 00:08:12 – On-premises SANs offered full flexibility to manage all metrics * 00:09:25 – Cloud autoscaling combined with software gives much more flexible storage * 00:13:27 – The on-demand nature of cloud works well for scaling SANs * 00:14:40 – New cloud features – NVMe, fast networking and NVMe/TCP have enabled solutions * 00:17:19 – What is NVMe/TCP? * 00:20:50 – What challenges are there in delivering a SAN on public cloud instances? * 00:24:03 – Cloud providers optimise for their system, not for your application * 00:25:02 – What operating system issues exist when building a virtual SAN? * 00:28:27 – Userspace operation requires a different programming strategy * 00:31:00 – NUMA awareness is essential, even in the public cloud * 00:32:46 – Each new instance type requires retesting and validation * 00:35:53 – What is the Lightbits solution and how is it deployed? * 00:37:00 – NVMe cloud drives are ephemeral * 00:41:35 – Snapshots work differently in the public cloud * 00:44:12 – Is Lightbits dedicated or HCI? * 00:46:32 – How is the solution consumed? * 00:47:29 – Azure offers management application capability * 00:50:36 – Wrap up

Related Podcasts & Blogs* Is the Public Cloud Becoming More Reliable? * Zesty Optimises AWS EC2 EBS Storage * Storage QoS In The Cloud * #97 – Building Storage Using NVMe/TCP with Kam Eshghi from Lightbits Labs * #121 – NVMe 1.4 Deep Dive Part II with J Metz * #210 – Building SANs in the Cloud

View Details

In this episode, Chris meets with Pure Storage co-founder John Colgrove (aka Coz) to discuss where the future of data storage lies in the enterprise. This recording was made at the October 2023 Pure//Accelerate event in London and has a little bit of microphone noise at the beginning (apologies!). However, keep listening for some great insights.

In this conversation, Coz explains how Pure Storage has put efficiency at the core of design and product evolution. However, this doesn’t just mean an increase in media capacity, but a vastly improved density and power consumption story. Gains are also being made in software, with continuous improvements in erasure coding overhead and data reduction techniques like compression.

Where will the future lie? NAND flash has many years to go yet, replacing the HDD entirely in the enterprise, if Pure Storage is to be believed. In the meantime, materials science improvements will continue to drive costs down and capacities up. Exactly how far those trends will continue remains to be seen, but we can guarantee that vendors like Pure Storage will be focused on continuous delivery of value to customers.

Elapsed Time: 00:20:36

Timeline* 00:00:00 – Intros * 00:02:55 – Change is in place in the storage industry * 00:03:22 – Power budgets are determining technology choices * 00:06:02 – Customers want greater efficiency * 00:08:35 – New chassis will allow more modules in the same physical space * 00:09:54 – A modern SSD has the capability of an entire rack of storage from 2000 * 00:11:31 – Can we envisage an 11PB flash module? * 00:12:24 – Gains aren’t always about increases, but about reductions * 00:13:02 – Pure Storage merges engineering, software and commercials * 00:15:44 – Materials science is driving the improvements in computing * 00:17:39 – New technologies will appear to replace flash * 00:18:20 – Optane was killed off by economics * 00:20:03 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 251 – Modernising Storage as a Service with Prakash Darji * Storage Unpacked 248 – FlashArray R4 Announcements from Pure Accelerate 2023 * Storage Unpacked 245 – Design Strategies fro 300TB Flash Drives with Shawn Rosemarin * Pure Storage announces Disaster Recovery-as-a-Service, SLA guarantees and operational cost rebates * Pure Storage: No New Hard Drives WIll be Sold by 2028 – Really?

View Details

In this episode, Chris is in conversation with Prakash Darji, VP and GM of the Digital Experience Business Unit at Pure Storage. The company has just announced a new Paid Power & Rack Space offering, new SLAs for service delivery and Pure Protect //DRaaS a new disaster recovery solution delivered as a service.

New customers of Pure Storage with Evergreen//One and Evergreen//Flex will be eligible for rebates on power and rack space costs over the term of their subscription. This is returned as cash or a discount on future purchases. New SLAs have been introduced that cover data durability (Zero Data Loss) and no forklift upgrades (No Data Migration).

Pure Protect //DRaaS is a new offering that provides disaster recovery as a service, from on-premises VMware vSphere deployments to the AWS public cloud and back again. The overall message is the transformation of Pure Storage’s business to be more service-focused and encompass all aspects of storage and data.

Elapsed Time: 00:36:17

Timeline* 00:00:00 – Intros * 00:01:15 – Announcements – Paid Power & Rack Space * 00:02:00 – New SLAs, Zero Data Loss, No Data Migration * 00:02:25 – New DR as a Service – Pure Protect //DRaaS * 00:03:00 – What is Pure’s SaaS Strategy? * 00:04:30 – Hardware tends to age (and get worse) over time * 00:04:45 – Pure offers both Hardware as a Service and Software as a Service * 00:05:30 – Legacy storage deployments are like a phoenix reborn * 00:08:10 – Why would Pure Storage pay customers’ data centre bills? * 00:11:00 – Costs were estimated with partners (Equinix, Digital Realty) and the IEA * 00:13:45 – Infrastructure TCO is important today, due to power costs * 00:15:00 – New SLAs – No Data Migration Guarantee – No Forklift Upgrades * 00:23:30 – Pure Protect //DRaaS – a slight diversion for Pure Storage * 00:26:15 – //DRaaS will protect VMware instances to AWS (and back) * 00:27:45 – DR used to be based on physical storage replication * 00:29:30 – Future Pure Protect solutions could use storage natively * 00:31:15 – Storage DR was never about the storage replication! * 00:33:00 – Pure1 is solidifying as the focus of Pure Storage solutions * 00:34:30 – When will the new features be available? * 00:35:30 – Wrap Up

Related Podcasts & Blogs Pure Storage Announces Disaster Recovery-as-a-Service, SLA guarantees and operational cost rebates * X-Ray: Pure Storage, Inc. * Pure Storage Introduces a Revamped Evergreen Programme * #185 – Pure-as-a-Service 2.0*

View Details

In this episode Chris gets an update on new announcements from Infinidat, including SSA Express and higher-capacity SSA II systems. SSA Express is a new feature of the InfiniBox “classic” platform that allows customers to implement what looks like a virtual SSA (all-flash array) within a hybrid InfiniBox. The solution effectively pins volumes in SSD rather than allowing the data to cascade to the HDD layer, with the benefit of SSA performance for no additional cost on 95% of hardware in the field.

Infinidat has also introduced a new SSA II, the F4316T with up to 6.6PB of effective capacity, doubling the previous maximum. The F4304T has been discontinued, so to meet the requirements of customers to scale from small to large systems, the existing F4308T and new F4316T can be initially deployed at 60%, 80% and 100% full. This scale-up option gives customers choice on how to deploy and extend all-flash implementations.

More information on the new offerings can be found at https://www.infinidat.com/ and more details on Infinidat at our Infinidat Microsite.

Elapsed Time: 00:33:51

Timeline* 00:00:00 – Intros * 00:01:00 – Quick update on announcements this year * 00:03:30 – Vendors need to be offering value-add * 00:04:25 – SSA Express is a software upgrade to add an embedded all-flash array * 00:05:30 – SSA gains expansion and scale-up capabilities * 00:06:40 – What is the SSA (and SSA II)? * 00:08:45 – SSA Express is a free software upgrade * 00:10:55 – SSA Express is similar to pinned volumes * 00:13:20 – Infinidat will validate the suitability for SSA Express * 00:14:50 – Most vendors don’t retro-fit a feature across all historical platforms * 00:16:55 – InfiniBox is generally deployed fully configured * 00:18:20 – SSA now offers double capacity with F4316T * 00:24:05 – The new scalability allows customers to grow capacity more easily * 00:25:35 – Partially populated SSA II still runs at 100% performance * 00:28:10 – Value add for the customer is important in the current market * 00:31:35 – Wrap Up

Related Podcasts & Blogs* Storage Unpacked 247 – Infinidat announces InfuzeOS Cloud Edition and InfiniSafe Cyber Detection * Storage Unpacked 231 – Introducing Infinidat InfiniBox SSA II * Storage Unpacked 227 – Infinidat InfiniGuard Enhancements with Eric Herzog * Infinidat Microsite * The Quiet Success of Infinidat

View Details

In this episode, Chris talks to Derek Dicker, CEO at Nyriad about the UltraIO storage array. Nyriad has developed a new storage architecture using GPUs that accelerate the calculations needed to store data using erasure coding. This enables UltraIO to implement system-wide data protection using erasure coding at the block level.

In contrast to most storage vendors in the market today, the UltraIO platform uses hard disk drives, with a GPU to process data ingested by the system, while data is presented back through the CPU route. This dual processor architecture enables Nyriad to deliver a product with 20GB/s of throughput, scale to multiple petabytes of capacity and provide dynamic data protection defined by the customer.

Nyriad sees UltraIO being used across four industries – HPC, Media & Entertainment, Backup and Recovery, and Active Archive. Essentially the solution excels at handling large volumes of unstructured data that needs high throughput processing.

Learn more about Nyriad, the origins of the solution with the Square Kilometre Array and customer examples at https://www.nyriad.io/

Elapsed Time: 00:32:28

Timeline* 00:00:00 – Intros * 00:01:40 – UltraIO was introduced in 2022 * 00:02:25 – Why is UltraIO different to traditional storage systems? * 00:03:30 – GPUs can be used within data storage systems * 00:04:10 – The Square Kilometre Array was an early customer * 00:06:15 – UltraIO fits a specific set of requirements around data ingestion throughput * 00:06:55 – UltraIO uses hard disk drives and erasure coding * 00:08:00 – Ingested data is processed via GPU, then accessed by CPU * 00:10:00 – Erasure coding allows customer-based resiliency settings * 00:12:00 – The hardware for UltraIO uses standardised off the shelf hardware * 00:14:50 – What markets does UltraIO fit? (HPC, M&E, Backup/Recovery & Active Archive) * 00:16:15 – The UltraIO architecture has strong sustainability characteristics * 00:18:45 – Most vendors have moved away from HDDs * 00:23:00 – Digital Image replaced three systems with an UltraIO * 00:24:20 – Don’t keep data forever! * 00:26:35 – UltraIO helped Digital Glue deliver a media asset management solution * 00:27:30 – System capacities are from one to three petabytes raw * 00:29:15 – Nyriad works through the channel * 00:31:00 – Wrap Up

View Details

In this episode, Chris talks to Dan Kogan from Pure Storage about the upgraded FlashArray R4 systems and new FlashArray//E platform.

View Details

In this episode, Chris discusses the latest announcements from Infinidat with CMO Eric Herzog, including a cloud version of the Infinidat platform and new cyber detection capabilities. Infinidat has released a version of the InfuzeOS that runs on the public cloud, marketed as InfuzeOS Cloud Edition. The software is a fully functional version of the same software running on all Infinidat systems, although at launch the implementation will operate as a single node. Customers can use the solution to replicate data into the public cloud and access it for a wide range of use cases, including DR, test/development work and capacity bursting.

Also announced is a new InfiniSafe feature called Cyber Detection. As the issue of ransomware continues to affect companies, faster and earlier recovery techniques are required. Cyber Detection enables Infinidat customers to detect potential malware in immutable snapshots using AI-based scanning techniques that provide feedback on the likelihood data contains ransomware software.

InfuzeOS is available today, while Cyber Detection will be released in 2H2023. For more information see https://www.infinidat.com/.

Elapsed Time: 00:36:07

Timeline* 00:00:00 – Intros * 00:01:30 – Infinidat and Eric have been winning awards * 00:03:00 – Infinidat announces InfuzeOS Cloud Edition * 00:04:20 – Also new, InfiniSafe Cyber Protection * 00:05:30 – InfuzeOS is the single software across all Infiniboxes and variants * 00:07:30 – Why is a single storage operating system so important? * 00:10:00 – Cloud edition enables customers to do DR, testing and other use cases * 00:11:05 – Homogeneous storage provides native and efficient data replication * 00:13:30 – Cloud Edition is AWS only at launch, other clouds to follow * 00:13:56 – Cloud Edition will be a single virtual instance today * 00:16:15 – Data persistence in the cloud is challenging to achieve * 00:16:50 – There will be a nominal fee for InfuzeOS Cloud Edition * 00:18:20 – Cyber detection is needed as the next defence against ransomware * 00:22:00 – Infinidat already does rapid data recovery from snapshots * 00:28:00 – Infinidat is shifting the ransomware detection to the left * 00:29:44 – AI determines whether content could contain malware * 00:32:30 – Customers can choose which data to scan * 00:34:10 – InfiniSafe Cyber Detection will be available in 2H2023 * 00:35:00 – Wrap Up

Related Podcasts & Blogs* #231 – Introducing Infinidat InfiniBox SSA II * #227 – Infinidat InfiniGuard Enhancements with Eric Herzog * #202 – Enterprise Storage Consolidation with Phil Bullinger from Infinidat * The Quiet Success of Infinidat

View Details

In this episode, Chris has a wide-ranging conversation with Fermyon co-founder and CEO, Matt Butcher. The main essence of the discussion is to look at serverless technologies and how storage features are integrated into that ecosystem. However, as you will hear, the content includes everything from Mainframes to Microsoft, and Kubernetes to WebAssembly.

Serverless is a technology that provides a “request/response” type operation for code execution. The request trigger could pass data in, or be requesting a data query. In any event, a serverless framework needs to offer security, multi-tenancy and stateful retention of data. Modern stateless environments are using WebAssembly (also known as WASM) to provide a secure and scalable runtime into which features are implemented that offer abstraction from the underlying platform.

Fermyon has created a developer framework called Spin, which now includes a key/value database implementation. You can find out more here – including trying the ecosystem out for yourself.

We expect to see much greater use of serverless technologies in the future, as the ecosystem matures and develops.

For more information on Fermyon, check out the company website.

Elapsed time: 00:53:00

Timeline* 00:00:00 – Intros * 00:02:45 – Microsoft and Open Source….hmm * 00:03:50 – WebAssembly (WASM) is set to be a leading area of technology * 00:04:30 – How would Matt describe “serverless”? * 00:07:30 – Serverless has lots of parallels to web server technology * 00:10:40 – Perl and PHP provided extensions to early GCI solutions * 00:14:50 – Mainframe had the multi-tenant security model * 00:15:30 – Microsoft derailed the mainframe * 00:17:55 – Security is an important aspect of serverless * 00:23:00 – What is the transactional nature of serverless? * 00:25:10 – Replicas cost money – but don’t have the same overhead in serverless * 00:27:30 – Billing should be called accounting * 00:28:15 – Lambda executes 10 trillion serverless functions per month * 00:30:30 – Features like networking and storage are essential serverless features * 00:33:15 – Let developers deal with state – right? * 00:36:45 – Application-based state is like keeping spinning plates in play * 00:40:05 – WASM provides a framework to run code without additional coding * 00:42:30 – Version 2.0 frameworks should enable plugin features * 00:43:50 – All storage (including databases) are just a function call away * 00:47:25 – Any implementation needs to balance development and operations * 00:50:30 – What is Fermyon building? Spin and Fermyon Cloud * 00:51:50 – Wrap Up

Matt’s BioMatt Butcher is the co-founder and CEO of Fermyon, the serverless WebAssembly company. Matt is one of the original creators of Helm, Brigade, CNAB, OAM, Glide and Krustlet. He has written and co-written many books, including “Learning Helm” and “Go in Practice.” He is a co-creator of the “Illustrated Children’s Guide to Kubernetes” series. These days, he works mostly on WebAssembly projects such as Spin, Fermyon Cloud and Bartholomew. He holds a Ph.D. in Philosophy and lives in Colorado.

Related Podcasts & Blogs* #200 – Virtualisation, Containers, Serverless and Data * #182 – FaunaDB – Client Serverless Computing * What’s Old is New at KubeCon * Databases – The Fourth (Storage) Protocol?

View Details

In this episode, Chris talks to Shawn Rosemarin (VP, R&D, Customer Engineering) from Pure Storage about the evolution towards 300TB direct flash modules, the custom-designed SSDs used in FlashArray and FlashBlade. Pure Storage has stated an intention to deliver 300TB modules by 2026. With only three years to achieve that goal, how will the company move from today’s 52TB maximum capacity to 6x the amount of storage in a similar footprint?

There are two key technologies at play here. Firstly, Pure Storage designs and manufactures custom SSDs (DFMs) rather than use commodity devices. This approach offers much greater control over the operation of NAND and other aspects such as DRAM usage and power draw. Second is the use of Purity, the Flash platform operating system, which directly controls the placement of data across NAND.

As DFM flash capacities continue to grow over the next three years, the difference between future HDDs and commodity SSDs will widen. Pure Storage hopes the ability to drastically increase capacity in the same footprint will offer customers a way to cope with environmental costs, the cost of data centre space and meet the increasing storage demands of technologies such as AI.

Here are links to the “better science” blog posts mentioned in the discussion:

https://blog.purestorage.com/products/better-science-volume-1-hardware-and-software-co-design-with-directflash/
https://blog.purestorage.com/perspectives/better-science-volume-2-maps-metadata-and-the-pyramid/

Here’s the link to our Pure Storage Microsite.

Elapsed Time: 00:46:18

Timeline* 00:00:00 – Intros * 00:02:20 – Pure intends to reach 300TB DFMs by 2026 * 00:03:40 – Why do we not have all-flash data centres? * 00:05:50 – Enterprise drives developed from consumer NAND usage * 00:06:35 – HDDs are 67 years old * 00:08:40 – 100TB HDDs by 2030? * 00:09:25 – HDD and SSD device capacity is diverging * 00:11:00 – Environmental aspects will affect adoption – such as power * 00:12:10 – Data centre growth is being restricted by power availability * 00:13:50 – there’s also an ESG aspect, what do businesses need to do? * 00:15:50 – How will Pure Storage get to 300TB drives? * 00:18:00 – Exploiting NAND efficiently is challenging to deliver * 00:22:15 – Firmware on standard drives is rarely (if ever updated) * 00:23:10 – Pure Storage can dynamically amend flash management algorithms * 00:24:25 – DRAM is used to obfuscate issues with flash management * 00:25:35 – Purity uses 1/40th of the amount of DRAM compared to standard SSDs * 00:27:40 – QLC NAND is progressively harder to manage than previous generations * 00:30:05 – How will reliability be managed at 300TB per drive? * 00:31:35 – How repairable are current (and future) SSDs? * 00:33:50 – As the biggest consumers of HDDs, how will the public cloud become all-flash? * 00:36:40 – The public cloud will move towards managing raw NAND * 00:38:12 – As divergence in capacity continues, businesses are creating greater technical debt * 00:41:10 – Traditional SSDs will still not keep up with DFMs * 00:43:30 – Pure Storage will reach 300TB in product release steps * 00:44:40 – Wrap Up

Related Podcasts & BlogsThe following podcasts and blogs are related to this podcast content.

Storage Unpacked 243 – Introducing FlashBlade//E
234 – Introducing Pure Storage FlashBlade//S and Evergreen//Flex with CTO Rob Lee
232 – Building Power-Efficient Storage Systems with Justin Emerson from Pure Storage
Pure Storage Announces FlashBlade//E
Pure Storage Announces FlashBlade//S
Dude, Here’s Your 300TB Flash Drive!

View Details

In this episode, Chris catches up with Omer Asad (SVP & GM) from HPE and Phil Manez (Director of Channel Sales) from VAST Data to discuss the recent announcement of HPE GreenLake for File Storage, powered by VAST Data’s Universal Storage platform. HPE and VAST Data have partnered to bring a new scale-out file services solution to market. HPE has developed a new platform of storage hardware, branded Alletra MP. The “file persona” is powered by VAST Data Universal Storage, marketed, sold and supported by HPE.

GreenLake for File Storage (and Block Storage sibling) represents a change in strategy for HPE, as software and hardware are licensed separately. This approach aligns with the direction adopted by VAST Data when Gemini was first announced in April 2021. From the perspective of VAST Data, the HPE partnership provides access to a broader market of customers, with an entry point as small as 250TB of unstructured data storage capacity.

Learn more about VAST Data at https://vastdata.com/ (or see the related links later in these notes).

More details on HPE GreenLake is available here – https://www.hpe.com/us/en/greenlake.html

Finally, here’s a link to the HPE GreenLake Day for storage event Omer mentions in the discussion.

Elapsed Time: 00:36:30

Timeline* 00:00:00 – Intros * 00:02:00 – HPE introduces GreenLake for File Storage * 00:03:35 – VAST is providing the storage component of the package * 00:04:40 – HPE has introduced new common hardware – Alletra MP * 00:06:30 – VAST Data’s Universal Storage aligned with HPE’s storage vision * 00:08:00 – Alletra MP will enable “personas” or protocol specific architectures * 00:09:30 – HPE storage licensing has evolved to separate hardware and software * 00:11:15 – Alletra MP enables customers to buy “storage hardware by the pallet” * 00:12:30 – VAST Gemini is the model of license disaggregation used by HPE * 00:15:30 – HPE’s servers bring supply chain and support benefits and reach * 00:18:15 – GreenLake storage servers are managed by the GreenLake infrastructure * 00:19:30 – Who supports the new platform? * 00:25:15 – VAST has helped HPE with third party InfoSight integrations * 00:27:30 – How will existing customers be managed? * 00:31:00 – What use cases does HPE expect with GreenLake for File Storage? * 00:32:15 – Customers can start at 250TB or 500TB * 00:35:45 – Wrap Up

Related Links* #233 – Introduction to VAST Data Ceres Hardware Platform with Jeff Denworth * #226 – VAST Data Universal Storage for Data Protection * #216 – The Data Explosion with VAST Data CEO Renen Hallak * #106 – Introduction to VAST Data (Part II) with Howard Marks * #105 – Introduction to VAST Data (Part I) with Howard Marks * #148 – Unpacking HPE’s Storage Strategy * VAST Data has Vast Ambitions * VAST Data goes software-only – what does this mean for customers? * VAST Data launches with new scale-out storage platform * HPE announces new GreenLake storage solutions with new hardware

View Details

In this episode, Chris is in conversation with Pure Storage International CTO, Alex McMullan, discussing the announcement of FlashBlade//E. The new FlashBlade//E platform extends the current offerings, targeting what the company terms the four “E’s” – environmental, economical, effortless and everlasting. Essentially, FlashBlade//E is a highly scalable version of the FlashBlade//S (which we discussed in June 2022). In the new platform, Pure offers both compute/storage and storage-only chassis that start at a capacity of 4PB and increase in 2PB increments.

The big difference between //S and //E models is the cost. Pure Storage has announced a target price of $0.20/GB raw, which is much lower than existing all-flash systems and is closely comparable with consumer SSD pricing. As a result, the FlashBlade//E is aimed at customers looking to retire HDD-based systems while gaining the benefit of random-access performance offered by all-flash systems.

There are some interesting aspects to this discussion. Alex explains that Pure Storage has a roadmap to 300TB direct flash modules, beating anything else on the market today. In parallel, the //E system aims to address the power consumption challenges of modern data centres, initially delivering approximately 1W per TB consumption (compared to 2W per TB today) with a target of milliwatt-level consumption per TB in the future. The third leg to this sustainability discussion is the significant reduction in e-waste compared to hard drives. All of these features are focused on helping customers deliver to their own sustainability targets.

FlashBlade//E is available for internal customer ordering and expected to ship within 30 days.

Elapsed Time: 00:34:49

Timeline* 00:00:00 – Intros * 00:01:00 – New announcement – FlashBlade//E * 00:01:50 – Data growth has exploded, is this sustainable? * 00:02:35 – FlashBlade//E is a capacity-optimised system * 00:03:45 – Data usage is changing, more data needs to be online * 00:04:30 – Sustainability – HDDs are hard to reuse, may get recycled * 00:06:10 – Pure is moving toward 300TB drives within a few years * 00:06:55 – Panorama (UK current affairs series) highlighted data centre impacts * 00:09:40 – Does the TCO work to replace HDD with all-flash? * 00:11:25 – Power costs have significantly affected TCO calculations * 00:12:00 – We also need to consider Total Cost of Value and Flexibility * 00:13:40 – Pure has 1 exabyte deployed at Meta, good learning environment * 00:14:35 – What does the FlashBlade//E system look like? * 00:16:15 – The //E model uses architecture ideas from FlashArray * 00:18:10 – FlashBlade//E evolves the issues with fixed node and blade capacities * 00:19:15 – 4PB (raw) is the entry point with 48TB drives, 300TB in the future * 00:21:30 – AFR rates is much lower on Pure custom SSDs compared to generic products * 00:23:25 – FlashBlade//E has two chassis types – how will performance change? * 00:27:00 – How does FlashBlade//E fit into the Evergreen consumption models? * 00:28:35 – How will data mobility be achieved between FlashBlade systems? * 00:30:45 – How does FlashBlade//E address our initial sustainability discussion? * 00:33:20 – FlashBlade//E is available within 30 days * 00:33:55 – Wrap Up

Related Podcasts & BlogsWe have many existing blogs and podcasts that discuss the evolution of the FlashBlade and FlashArray portfolios.

234 – Introducing Pure Storage FlashBlade//S and Evergreen//Flex with CTO Rob Lee
232 – Building Power-Efficient Storage Systems with Justin Emerson from Pure Storage
209 – Discovering Unified Fast File and Object With Pure Storage
133 – FlashArray//C Deeper Dive with Pete Kirkpatrick
Soundbytes #009 – FlashArray Update with Ivan Iannaccone at Pure Accelerate
Soundbytes #008 – FlashBlade 2.0 with Rob Lee at Pure Accelerate
Pure Storage Announces FlashBlade//S
Pure Storage Introduces FlashArray//XL
Pure Storage FlashArray//C

View Details

In this episode, Chris talks to Matt Hamilton, Developer Advocate at Protocol Labs, about the Filecoin decentralised storage network. Filecoin currently stores around 15 exabytes of data globally, across 4000 storage providers. The ecosystem enables end-users to contract for storage capacity and store data using the Filecoin marketplace platform. Customers and providers interact with FIL, the Filecoin currency.

Matt explains how Filecoin implements data immutability through a series of cryptographic proofs, including proof of capacity, proof of replication and proof of space/time. The platform enables 3rd party service providers to layer on additional features, such as consolidation for smaller workloads, caching and content delivery networks.

Most recently, Matt has been working on FVM, the Filecoin Virtual Machine, a system to enable smart contracts. This solution enables end users to build policies or actionable workflow to optimise storage usage or ensure continued storage of data.

To find out more, follow the following links:

Main Filecoin website – filecoin.io
Filecoin Virtual machine – https://fvm.filecoin.io/
Spacewarp – https://spacewarp.fvm.dev/

Elapsed Time: 00:39:55

Timeline* 00:00:00 – Intros * 00:01:15 – What is Filecoin? * 00:03:20 – Filecoin is a enterprise-class system, with around 4000 storage providers * 00:05:00 – Filecoin enables easy replication and “proof of storage” called proof of space/time * 00:06:45 – Providers puts up collateral, clients put up a fee, both kept in escrow * 00:07:30 – Proof of replication – shows the provider received and stored the data * 00:08:15 – Data isn’t stored on blockchain, only the audit and tracking data * 00:09:45 – Encryption is a layer above Filecoin * 00:10:30 – Storage aggregators consolidate smaller file and object into larger blocks * 00:11:35 – Filecoin is archival/cold storage with the ability to layer on additional services * 00:14:30 – Filecoin doesn’t waste the “proof of capability” resource * 00:16:01 – How does the dynamic marketplace of Filecoin work? * 00:20:20 – How does data move in and out of the Filecoin network? * 00:21:02 – Protocol Labs is working a CDN solution called Saturn * 00:21:30 – Filecoin Virtual Machine will enable smart contracts * 00:24:50 – Protocol Labs is working on bringing compute to data * 00:27:50 – End users can join the network as a CDN through Saturn * 00:28:45 – IPFS – Interplanetary File System! * 00:30:00 – Filecoin uses a content addressable ID * 00:31:10 – What are the typical users and use cases of Filecoin? * 00:32:55 – Many NFT platforms store content on Filecoin * 00:35:00 – Filecoin can be used to prove immutability in the future * 00:37:40 – Where can people learn more? * 00:38:30 – Wrap Up

View Details

In this podcast episode, Chris reviews the results from a recent project with Ondat, which examines the relative efficiency of container-native storage solutions running common database platforms on Kubernetes. Ondat CTO Alex Chircop and Customer Success Architect Chris Milsted join the conversation.

The work expands on an initial project published in January 2021 that performed synthetic benchmark testing against a range of commercial and open-source container-native storage platforms, including Ondat. This new report provides improved real-world insight, demonstrating the relative performance of MongoDB, PostgreSQL and Redis solutions.

The testing shows how Ondat, Portworx, OpenEBS, Longhorn and Rook/Ceph perform when running database performance testing software. The results measure transactional output including latency and throughput. An extra aspect of this testing is the additional data covering outliers; effectively understanding how well the solutions manage to deliver predictable I/O responses.

In this episode, both Alex and Chris explain the reasons why testing is so important for customers. The relative efficiency of container-native storage solutions varies widely, which can result in additional hardware expense for customers. This is of critical importance both on-premises and in the public cloud.

To access the report, visit https://www.ondat.io/benchmarking (registration required). The original synthetic benchmark testing can be found here

https://www.architecting.it/product/performance-benchmarking-cloud-native-storage-solutions-for-kubernetes-ebook/

Elapsed Time: 00:47:53

Timeline* 00:00:00 – Intros * 00:01:00 – Chris is the guest this time! (almost) * 00:02:15 – The first benchmark looked at synthetic testing (fio) * 00:03:30 – What were first thoughts and opinions? * 00:05:20 – What are the performance metrics to be considering? * 00:07:30 – Device evolution is giving us cheaper and faster storage * 00:08:15 – Kubernetes storage efficiency should match application density * 00:10:00 – Performance analysis provides application-specific insights * 00:11:15 – Operations per second is a better measure of application performance * 00:12:30 – What happens with 32TB and 100TB drives? * 00:15:30 – Networking, scalability and failures means there’s a lot of things to manage * 00:16:45 – Strong consistency and checksums are table stakes for CNS * 00:19:30 – What did the performance analysis report show? * 00:21:30 – Creating a fair test environment is a challenging process * 00:23:15 – This report was able to look at performance outliers and predictability * 00:26:15 – Ondat spends significant effort in performance tuning their solution * 00:28:45 – Performance management is an ongoing challenge as infrastructure evolves * 00:31:00 – CNS will act as a translation layer for applications and storage * 00:37:00 – What can we learn from the output of this testing report? * 00:39:45 – What did Chris E find interesting in the report results? * 00:42:15 – What could we test next? * 00:47:00 – Wrap Up

View Details

In this episode, Chris is joined by Phison CTO Sebastien Jean to discuss the transition from HDD to SSD. As pricing for the two media types comes close to reaching parity, this conversation looks at how the two media are evolving and how NAND flash SSDs are set to become the de facto choice for the enterprise data centre. The transition to an all-flash data centre has been predicted for some time, however end game could be on the horizon as soon as 2025 or 2026.

The Phison blog discussed by Sebastien can be found here - https://phisonblog.com/

Elapsed Time: 00:44:01

Timeline

00:00:00 - Intros

00:02:45 - Predictions of the death of the hard drive are overdone

00:04:15 - Enterprise SSD prices continue to drop - but don’t HDDs do that too?

00:05:30 - 2020 was the cutover year for units produced (SSD vs HDD)

00:06:45 - HDD capacity growth has slowed

00:07:50 - SSD growth is being achieved with 3D techniques

00:09:15 - Floors could be used to increase layer count

00:11:00 - Latency is increasing in newer flash designs

00:13:05 - What is the “acceptable” limit for SSD capacity?

00:14:15 - Dense SSDs will be the norm with NVMe connectivity

00:15:30 - New form factors will drive greater drive capacities

00:17:00 - Is $/GB too simple a measure?

00:19:40 - SSD rebuilds are quick, avoiding the need for RAID-6

00:20:45 - SSD failure is not predictable

00:25:00 - Are HDD shipping costs relevant?

00:26:30 - How are the hyper-scalers using HDDs and driving change?

00:29:15 - Where do we go next - PLC?

00:32:00 - Could an entire SSD be dynamic between SLC and QLC?

00:35:45 - What about new technology Optane 4.0 or MRAM?

00:40:30 - Look out for CXL-enabled SSDs

00:42:40 - 2Tb NAND dies will bring HDD parity

View Details

The PCI-SIG recently announced the next generation of PCI Express – 7.0 – will be released in 2025. We’ve only just heard about PCIe 6.0 (see this recent podcast), so it’s interesting to see that the working group is already committed to extending the performance and capabilities of PCIe further into the future. Chris caught […]

The post #239 – Unpacking the details of PCI Express 7.0 with Al Yanes appeared first on Storage Unpacked.

View Details

Once again Chris is joined by Rick Kutcipal, board member with the SCSI Trade Association. It’s been four years since episode #74 when Rick provided an update on SAS 24Gb. This time, Rick has details of new features in 24Gb+. What does the “plus” mean in this context? While the bus speed may not be […]

The post #238 – SAS 24GB+ Updates with Rick Kutcipal appeared first on Storage Unpacked.

View Details

In this episode, Chris chats to friend of the show Chris Mellor from Blocks and Files. Both Chris’s attended Flash Memory Summit this year, held annually in San Jose and in person for the first time since 2019. Two big technology developments stand out from this year’s event. Firstly CXL, which is current flavour of […]

The post #237 – Flash Memory Summit 2022 in Review with Chris Mellor appeared first on Storage Unpacked.

View Details

In this episode, recorded life at FMS 2022, Chris chats to IBM Fellow, Storage CTO and FlashSystem Architect, Andy Walls.

The post #236 – A FlashCore Module Primer with IBM Fellow & Storage CTO Andy Walls appeared first on Storage Unpacked.

View Details

In this episode, Chris is joined by IBM Storage CMO Scott Baker. The discussion takes place at Flash Memory Summit 2022 (FMS), being held in person in Santa Clara, USA. FlashCore Modules are IBM’s custom flash drives, developed from IP acquired with the acquisition of Texas Memory Systems. Scott is presenting at FMS, discussing how […]

The post #235 – FlashCore Futures with IBM CMO Scott Baker appeared first on Storage Unpacked.

View Details

In this episode, Chris talks to Rob Lee, CTO at Pure Storage about today’s announcement covering FlashBlade//S and Evergreen//Flex. FlashBlade//S is a new hardware platform that evolves the solution into a disaggregated architecture. The design enables compute and storage resources to be scaled independently, while incorporating existing DFMs (DirectFlash Modules) that are currently used in […]

The post #234 – Introducing Pure Storage FlashBlade//S and Evergreen//Flex with CTO Rob Lee (Sponsored) appeared first on Storage Unpacked.

View Details

In this episode, Chris is in conversation with VAST Data CMO, Jeff Denworth. The topic covers the recent announcement of Ceres, a new hardware platform combining ruler form-factor flash and BlueField DPUs. The new solution, which will eventually replace the current “Mavericks” D-boxes is a 1U system with flash, storage-class memory and network connectivity through […]

The post #233 – Introduction to the VAST Data Ceres Hardware Platform with Jeff Denworth (Sponsored) appeared first on Storage Unpacked.

View Details

In this episode, Chris talks to Justin Emerson from Pure Storage about modern design considerations for power-efficient storage. The cost of powering data centres has increased massively over the last 12 months, driven by the global economy and recovery from COVID-19. At a component level, servers are increasing their power usage, with faster processors and […]

The post #232 – Building Power-Efficient Storage Systems with Justin Emerson from Pure Storage (Sponsored) appeared first on Storage Unpacked.

View Details

In this episode, Chris talks to Eric Herzog, CMO at Infinidat about the announcement of SSA II, the second generation all-flash InfiniBox platform. The first generation SSA (solid state array) was developed to address a customer requirement that all long-tail latency I/Os must be delivered at all-flash levels of performance. The InfiniBox architecture is capable […]

The post #231 – Introducing Infinidat InfiniBox SSA II (Sponsored) appeared first on Storage Unpacked.

View Details

In this episode, Chris and Martin talk to Tim Walker and Mohamad El-Batal about the development of NVMe Hard Drives.

The post #230 – Seagate NVMe Hard Drives appeared first on Storage Unpacked.

View Details

In this episode, Chris and Martin talk to Al Yanes, President and chair of the PCI Sig, on the announcement of the PCIe 6.0 specification.

The post #229 – Exploring the PCIe 6.0 Specification with Al Yanes appeared first on Storage Unpacked.

View Details

In this episode, Chris talks to Murli Thirumale VP & GM of the Cloud Native Business Unit at Pure Storage about data agility.

The post #228 – Exploiting Data in the Third Wave of IT Agility with Murli Thirumale (Sponsored) appeared first on Storage Unpacked.

View Details

Infinidat has announced significant improvements to InfiniGuard. In this podcast, Eric Herzog, CMO updates Chris with the details.

The post #227 – Infinidat InfiniGuard Enhancements with Eric Herzog (Sponsored) appeared first on Storage Unpacked.

View Details

In this episode, Chris and Martin discuss data protection on a “vast” scale, with VAST Data VP of Data Protection, George Axberg. VAST has just announced a partnership with Commvault that will integrate Universal Storage with Commvault’s data management solutions. Further data protection vendor partnerships are expected in due course. While the idea of using […]

The post #226 – VAST Data Universal Storage for Data Protection (Sponsored) appeared first on Storage Unpacked.

View Details

Chris and Martin discuss the challenges of database sharding, scaling and performance with speedb CEO Adi Gelvan.

The post #225 – Accelerating Modern Databases with Speedb appeared first on Storage Unpacked.

View Details

Chris and Martin discuss FinOps and the process of public cloud resource optimisation with Aran Khanna from Archera

The post #224 – FinOps and Optimising Cloud Consumption appeared first on Storage Unpacked.

View Details

In this episode, Chris chats to Danny Abukalam, Product Engineering at SoftIron about the deployment of task-specific SDS solutions.

The post #223 – Task Specific Software-Defined Storage appeared first on Storage Unpacked.

View Details

In this second episode reviewing storage news of 2021, Eric Herzog, CMO at Infinidat joins the team to bring his opinion to the discussion.

The post #222 – End of the Year Show 2021 – Part Two appeared first on Storage Unpacked.

View Details

This week, the team take a review of 2021, looking at the events in storage and a lot more. In fact, this episode goes way off-piste in the conversation, so be prepared for public cloud, China, service models and storage unicorns.

The post #221 – End of the Year Show 2021 – Part One appeared first on Storage Unpacked.

View Details

Chris and Martin review the file system announcements coming out of AWS Reinvent event. Specifically FSx for OpenZFS and NetApp ONTAP.

The post #220 – AWS & Cloud File System Diversity appeared first on Storage Unpacked.

View Details

In this week's episode, Chris talks to Tony Mendoza from Spectra Logic about their recent ransomware attack and recovery.

The post #219 – Anatomy of a Ransomware Attack appeared first on Storage Unpacked.

View Details

This week, Chris is in conversation with Don Foster, Global Head of Sales Engineering at Commvault, looking at real-world customer challenges.

The post #218 – Real World Customer Data Protection Experiences with Don Foster (Sponsored) appeared first on Storage Unpacked.

View Details

Chris and Martin talk about the features and benefits of CXL, or Compute Express Link with CXL Chair Jim Pappas.

The post #217 – Introduction to CXL with Jim Pappas appeared first on Storage Unpacked.

View Details

This week, Chris chats to VAST Data CEO, Renen Hallak about the explosion in data and the challenges of analysing and managing huge quantities of unstructured content. VAST Data has built a scale-out unstructured object store that customers are now using to store petabytes of complex data. But what does complex mean, and how is […]

The post #216 – The Data Explosion with VAST Data CEO Renen Hallak appeared first on Storage Unpacked.

View Details

In this week’s episode, Chris and Martin discuss network booting and the ability to create completely stateless servers. The idea of SAN booting has been around for two decades, enabling both the operating system disk and data disks to be delivered from SAN storage. NVMe/TCP and other NVMe-oF protocols promise the ability to implement the […]

The post #215 – Stateless Compute Machines appeared first on Storage Unpacked.

View Details

This week, Chris and Martin discuss the announcement from CloudFlare of R2, a new object storage solution. With such a widely dispersed CDN network, could CloudFlare become a real competitor for S3, considering that the R2 platform will have no egress fees? This episode digs into the details, to look at what additional features object […]

The post #214 – Can CloudFlare R2 Disrupt AWS S3? appeared first on Storage Unpacked.

View Details

With KubeCon North America 2021 looming, Chris and Martin spend this week looking at data protection for Kubernetes-native environments. Vendors have started to introduce backup solutions, some working within the container environment itself, some outside as part of existing products. Exactly what should be backed up and how should these products work? Are we right […]

The post #213 – Kubernetes-Native Data Protection appeared first on Storage Unpacked.

View Details

In this week’s podcast, Chris and Martin get together to discuss the curious case of missing files. In this instance, the files weren’t actually missing, but according to an article in The Verge, students running a software simulation didn’t know how to determine their location. In a world where the current generation have grown up […]

The post #212 – File Not Found appeared first on Storage Unpacked.

View Details

In this week’s sponsored episode, Chris chats to Alex McMullan, CTO International at Pure Storage. The discussion covers Pure//Launch and today’s announcements of two SaaS solutions – Fusion and Portworx Data Services. Fusion is a new SaaS management plane for administering and optimising Pure Storage hardware and software products, including FlashArray, FlashBlade and Cloud Block Store. Portworx Data Services enables customers to deploy and operate managed databases using the PX-Enterprise platform. Together, these two solutions provide an insight into the next stage of evolution and growth for Pure Storage.

There’s a lot to unpack in this week’s podcast. Alex starts with some background on why customers are demanding a more cloud-like consumption model from vendors. The conversation then deep-dives into the details of Fusion and Portworx Data Services, finishing with a summary of the new solutions and how they highlight the next growth decade for the company.

You can learn more about today’s announcements through the following links:

https://www.purestorage.com/enable/modern-data-storage.html
https://www.purestorage.com/enable/pure-fusion.html
https://www.purestorage.com/enable/portworx/data-services.html

Elapsed Time: 00:52:17

Timeline * 00:00:00 – Intros * 00:01:46 – What is //Launch? * 00:02:45 – How are customers’ requirements evolving? * 00:05:00 – Cloud has changed the demand model for customers * 00:06:20 – What are Fusion and Portworx Data Services? * 00:08:06 – Both new offerings are SaaS * 00:08:42 – Digging down into Fusion * 00:11:00 – Fusion optimises the Pure Storage “fleet” of storage platforms * 00:14:15 – What features in Pure products enable better SRM? * 00:18:00 – Fusion uses Storage Classes and Protection Policies * 00:19:30 – How will Fusion mix multiple commercial business models? * 00:22:45 – What is the charging structure for Fusion? * 00:24:30 – What backend support work is required for Fusion? * 00:26:15 – Observability is a big part of Fusion * 00:28:40 – What is Portworx Data Services? * 00:31:45 – What will the Portworx Data Services “experience” be? * 00:36:30 – PDS enables “complex” storage deployments * 00:39:00 – Portworx Data Services uses a local management “Operator” * 00:40:40 – PDS will aid DBAs, not put them out of a job! * 00:42:15 – How will Pure1, Fusion and Portworx Data Services integrate? * 00:44:10 – Will Portworx Data Services work in private and public cloud? * 00:46:30 – Hardware is not going away? * 00:51:45 – Wrap Up

Transcript

Show More Speaker 1:
This is Storage Unpacked. Subscribe at storageunpacked.com.

Chris Evans:
This is Chris Evans recording the Storage Unpacked Podcast. And this week, it’s just me. Or rather, just me with a guest, actually. And I’m joined today by Alex McMullen from Pure Storage. Hi Alex, how are you?

Alex McMullen:
Hey, Chris. Good morning. Looking forward to the conversation.

Chris Evans:
Absolutely. So we’re here because today you’ve had a big announcement from the company, something that’s called //Launch on all the publicity that we’ve seen, which is a great name for me to try and remember every single time. However, it’s quite a big launch and we’re going to dig into that in a second and find out more. But before we do that, could you just give people a bit of background about yourself? And then we’ll chat about what you’re actually announcing today.

Alex McMullen:
Sure. Morning, everybody. My role at Pure is fairly an interesting one. I’ve been here now just on seven years. My job title as it stands is CTO international. And I represent the office of the CTO outside Mountain View along with a very talented team of folks in all of the theaters. And our really big push for this year has been to get the modern messaging out around Pure. We’re really focused on making sure that customers understand why Pure is doing what it’s doing, but also more importantly, why we believe that’s relevant in the hugely rapidly moving technology world that we love to be part of these days.

Chris Evans:
Brilliant. Today we’re here because there’s a big launch event. It’s not a product launch event. It’s actually been very interesting watching some of your pre-event media saying, this isn’t about new hardware or new boxes and things like that, it’s a different approach. So what exactly is Launch about and what are you, very briefly, going to launch or have you launched today?

Alex McMullen:
Yeah. So this is super Tuesday for us. It’s going to be quite interesting to see how the market wants to respond actually because traditionally, we’ve come up from that storage control plane, if you will, since 2009 and we’ve iterated our product line since then. But really, this is our first foray into the real disclosure of Pure’s actual capabilities. So this is very much a focus for us on software as a service on modern cloud agility, but also really solving for the problems of the 21st century, which many of our previous competitors even have not quite gotten to that point yet. So this is our SaaS control plane for a different set of our product lines and we’ll get into those details as we go forward.

Chris Evans:
Fantastic. For me, that first piece, I think is quite interesting. The fact that we’re talking SaaS here and this seems to be a big part of the way that the market is changing in general. And I think this is worth touching on very briefly because it’s something that you must be seeing with customers all the time who want to try and evolve their business, not to be thinking about the management of infrastructure, which, who wants to spend all that time doing that? And more about allowing that to be translated into a service model. So what are you seeing out there in the market in terms of what people want and how this has evolved since, like you said, 2009, when the company first was founded?

Alex McMullen:
What we’ve seen in particular since, I guess the creation, even of the iPhone and the App Store back in 2008 is the rise in influence and capability of software developers in every company. And what that’s really driven is the appetite, the ambition for pretty much everyone who is in technology, and some who are not, to go faster, go further and to be more creative, more intuitive but also to give that cloud consumption experience as well as doing it in a more elegant, more simple way.

Alex McMullen:
That’s really what we’ve learned and that’s how Pure has a great track record in terms of not just predicting what customers want in future but also making big and bold bets in getting there in the first place. And of course, we have a track record in doing that in terms of starting with all FlashArrays and then putting data reduction on but more importantly, our Evergreen capabilities and also the concept of having a SaaS portal, which is Pure1, even back in 2014, where you could see the entire fleet status of every Pure device you owned.

Alex McMullen:
And we continue to iterate down since then, in terms of all the clever things we’ve done with creating our own MVME drives, which allows us to be much more agile with use of the medium and then that reusable [inaudible 00:04:16] capability with FlashBlade and FlashArray, of course. But really what you’ve noticed over that time is incremental capability has all being taking us down a path. And really now we’re starting to show the cloud-like economics and cloud-like capabilities that we have in play.

Alex McMullen:
We dealt with Pure as a service last year, utility consumption for the cloud. And now with the acquisition of Portworx, we’re able to solve for that not just in traditional apps but also in any control plane or any public cloud or any hyperscaler with the capabilities that Portworx brings. And that’s really what brings us back to the announcements for today.

Chris Evans:
I think the evolution of the industry is really interesting in the sense that cloud, and by that we mean probably public cloud more than anything else, has driven people to realize that they don’t really have to care about the underlying infrastructure. I think you look back to virtualization in that initial step where people thought, “Ooh, I’m not sure I really want to trust the virtualization layer will do what I need it to do.” Very quickly people were like, “Well, it works, why are we even bothered about this?”

Chris Evans:
Cloud has formed that similar approach for people where initially there was a lot of discussion about whether we could trust the client, but then fairly quickly people realized it was reliable, it was trustworthy. And then I think people question, “Well, why do I need to spend all my time managing this infrastructure when I could have a vendor or somebody else do it on my behalf?” And that, to me seems the story of where we’re headed today with the announcements, which I guess you’re about to tell us.

Alex McMullen:
Yes. I mean, I like that point as well. Of course, if you go back to even the days of RAID, Storage gets bad press for being slow and not particularly creative, but storage did virtualization before any of the other guys, you could probably argue maybe mainframe partitions. But I think if you look at that and going back to the early days of the work that Gibson did and Bianca Schroeder are falling on from that in terms of their liability side, we’ve been very good as an industry in terms of federation, virtualization, leading that technology charge but doing it very quietly. And that’s the ethos and the mindset that Pure has taken forward here. And that’s really what we want to talk about today overall. So how do you want to play this, Chris? Where should we start on all these whole feast?

Chris Evans:
The feast of announcements. Why don’t you just start by giving us a high level view of the main components that have been talked about today?

Alex McMullen:
So the two main things that we’re here to talk about today, firstly, is a SaaS product offering on our existing FlashArray/FlashBlade capabilities. And what that brings us really is a cloud-based portal, which allows you to manage, federate, maintain, monitor and also workload balance. All of your internal workloads within FlashArrays and FlashBlades, but also doing the same thing with Cloud Block Store instances in the public cloud.

Alex McMullen:
So really that’s a big change for us because what’s really happening here is you’re now seeing a control plane running in the public cloud, which is actively managing, monitoring and controlling infrastructure that’s running on-premises or in the public cloud. So that’s a very much a multifaceted from how traditionally we’ve looked at infrastructure, but what we’re doing here is giving that SaaS level experience for fast moving agile, but also very efficient infrastructure.

Chris Evans:
Okay. So that’s the first one.

Alex McMullen:
That’s one. And the second is, doing something even more ambitious on the containerized side. We’re very fortunate with Portworx in that it already had many of these, what I describe as a service capabilities because of its work with all the hyperscalers, but we’re taking that step even further with that same ambition of simplicity, cloud-like consumption and allowing again with a SaaS portal running in the public cloud, to support the ambition of any Kubernetes cluster to automatically deploy and manage a number of SQL and no SQL databases. So it was database as a service for Kubernetes applications. And again, that’s based on what we’ve seen in the market directionally, but also based on what customers are asking us for.

Chris Evans:
Just to qualify and just to make sure that was really clear, both of these are SaaS offerings. So there’s no “New” hardware products here or anything like that. This is all about SaaS, all about service delivery.

Alex McMullen:
Exactly. That’s the ethos. Of course, we’d love you to buy more stuff at the same time, but it’s quite happy to work with your existing deployments and that’s the key thing here. On the Portworx side, we want to remain Switzerland in that whole story. Portworx is software defined storage, and we’re not locking this into Pure platforms. We consciously made the investment decision with Portworx to continue to work with the wider industry and with white boxes and with the hyperscaler hardware.

Chris Evans:
Let’s dig down into Fusion first of all and understand a bit more about that. Now, the name itself gives it away somewhat, doesn’t it? Fusion, the bringing together of a solution that allows you to manage all of your existing, let’s call it storage layer products. I know in some of the presentations that you’ve used, you’re using the idea of a layer cake. I’d never heard of a layer cake by the way, until it was mentioned on the bake off a few years ago. But now I know what a layer cake is, but you’re sort of using this analogy of the layer cake or even just think about things like protocols, like networking, where they have layers already. And at the very bottom here are the physical components like the infrastructure like FlashArray, FlashBlade and Cloud Block Store sitting on the public cloud. So the idea here is to bring these together in a federated manner so that you can actually allow customers to, I guess, abstract, simplify. I’m probably putting words into your mouth. So maybe you could go into a bit more detail as to exactly how this would work.

Alex McMullen:
Yes. So what we’ve really been doing here and building up for the last few years is that series of layers, if you will, which if you want to look at the OSI model, it’s something similar. The lowest level functionality is the hardware platform and the operating system code that runs on that. And what we have been doing is adding those layers accretively, which give us actual layers of functionality over the last few years. Some hardware, some software, but really the big steps forward have been with Pure1 and now with the Fusion product line.

Alex McMullen:
Like many of us in the industry, some of our product naming strategies leave us looking a little perplexed, but I think Fusion is actually one of the ones we’ve got spot on. The way in which that brings that mental image of that molding, that bringing together, the automatic lovely engagement of all those different layers together is very much the focus here. It brings together the whole cloud operating model. It’s a SaaS portal. So there’s no infrastructure or no extra management, no VA to take care of, none of those kinds of things. What it really is doing is taking that one step further.

Alex McMullen:
Our product lines today already have great rest APIs, but we’ve built in those functions like sync replication with our active cluster product to move LUNs. We’ve been doing that for a few years and some customers do that with their own Python scripts today already. But what we’re talking about here is that Fusion layer that listens to the telemetry from all of the Arrays that are already coming back home or new ones as they come forward. And what it does is essentially based on rules that you as a customer want to define in terms of latency or capacity or performance or cost, it will then seek to optimize the deployment of all of those Arrays.

Alex McMullen:
And if we look at that from a high level, what we’re really doing is separating that layer cake in terms of the people that manage the infrastructure and taking that away in a separate layer from the people who want to consume the infrastructure. And what that means is, if you eat cake the way I do, you always eat the icing bit on the top and then get down to the good stuff afterwards. But what this means is, the personas who consume storage via software, whether that’s TBAs or whether that’s data scientists or engineers or researchers, they make a rest API call to get some capacity within their tenant space, because Fusion brings the same concepts as they’re used to having in the public cloud. So it brings you region availability zones, it brings you the policies and classes of the capabilities that are available.

Alex McMullen:
And there are several layers of obstruction in that so that Fusion brings you all the great things you want in terms of, I’m responsible for my own destiny, I’ve got my own piece of capacity that I get to consume how I want and I get billed on that basis depending on the usage profile for me. So it’s cloud-like personas, interacting with concepts that they’re familiar with, but also being done completely with on-premise infrastructure in the first phase and then public cloud, of course, as well.

Alex McMullen:
But all through that SaaS portal, I don’t want to call it AIOps because everybody goes, “Oh really? AIOps, here we go.” But it’s actually based on machine learning experience that we developed over the last few years within Pure1 and Pure1 Meta and giving that whole stratospheric view of, here’s my fleet from an infrastructure management point but more importantly as the developer or the consumer, here’s how I get to use my piece of capacity and here’s how it gets better, faster, cheaper over time. And if I want to move the policies or change the policies, that all magically happens behind the scenes going to the higher capacity platforms or new protocols as they’re added, et cetera. So that’s the 10,000 foot view, I guess, Chris, where do you want to go?

Chris Evans:
I’d just like to point out some really interesting things within this. First of all, probably 10, 15 years ago, I’ve worked at a large, let’s call it a large financial organization, probably one of the ones that is the largest in the US at the moment. And we spent a lot of time trying to work out how we were going to do deployments. And even then when we looked at some of the SRM tools that were floating around in those days, that were pretty average, most we’ve never really had any SRM tools that have really done the job properly. But most of the time it was fairly obvious that you had consumers who were sitting at, let’s call it, above the line, who wanted to actually just come in, use resources and needed an API that was very focused on their ability to actually request and use those resources.

Chris Evans:
But then you had people, let’s call it, below the line, who were the admins who were sitting there thinking, “Well, I’ve got to make sure there’s enough capacity that I’m deploying stuff when I need it, I need to be able to block out the Arrays that might be being decommissioned so people can’t keep allocating on them but I don’t want to do that in a manual fashion, I want to try and automate this.” And what you’ve described in the initial implementation here at Fusion is exactly that thing that we’ve been trying to do for probably 10 to 15 years and hasn’t been possible to do. So why do we think that we’re going to manage to do that now? What has changed within the product set? What’s allowed us to get to that point where in fact, these types of SRM tools are going to actually work.

Alex McMullen:
A number of things to touch on there, lots of good food for thought on that side. We don’t really see Fusion by the way as what you would call an SRM layer, this is more of a federation. I don’t want to get into Star Trek analogies this early, but I’m sure lots of people will be joining the dots there. Really what we’re talking about here is an intelligence layer, if you will, which allows a presentation layer for consumers and a management layer for the infrastructure managers, the data center owners, those kinds of folks too. So we’re not losing sight of that.

Alex McMullen:
What really jumps out here is that Fusion leverages a very high capability set of products in a very thoroughly implemented rest API, which themselves are concepts that we’re selling. I was used to, as we should have done. Back in the day things that popped up, we had various virtualization capabilities, everybody from Hitachi and VSP and USP, and IBM with SBC and the B-series. Even down to things like EMC ViPR as well, but they all tried to minimize and standardize the level of capability on the products they sat in front of.

Alex McMullen:
Fusion is the antithesis of that in terms of, it leverages the full capabilities of the platforms, hardware and software that it sits in front of. The capability to move LUNs around between Arrays without anybody noticing. The ability to move from one class of device to another, from FlashArray//X to C and to FlashBlade. Those are things that nobody else in the industry can do even today. So what Fusion really is doing is bringing in best of breed. Of course, we have a view on why that is, to the customer personas and allowing them simply to do what they want to do. It comes back to that simplicity messaging, the cloud-like messaging.

Alex McMullen:
Nobody wants to get into conversations about LUNs and zoning and the number of paths and port over subscription. Those are conversations are no relevant anymore in the public cloud ecosystem. So this is very much about the simplicity, the cloud-like consumption, and Fusion will continue to add those capabilities. We already have things like monthly billing, those kind of integrations with Prometheus and Grafana. So it is fully modern in terms of everything from monitoring to observability, which is a very much a growing trend as well there. So we look at Fusion really as the intelligence layer, rather than an SRM, because I think in terms of ambition is much more broad and much more aggressive than that.

Chris Evans:
The term SRM is massively loaded from history-

Alex McMullen:
Oh, yes.

Chris Evans:
…in terms of the way that we think about it. And in some respects, that was a deliberate, let’s call it a deliberate ploy to use the term SRM there because people will be familiar with that and not necessarily be thinking how that translates into something higher. Clearly one of the things that we’re definitely very familiar with inside the cloud or I think even just through things like virtualization is, the concept of policy. And when you look at policy and when you look at things like storage catalogs or any sort of catalog that you build for resource consumption, you start realizing that you actually need to abstract away from the physical infrastructure that sits underneath and start talking about things that are more application-focused and business-focused. So how does that fit in with the way that Fusion is been deployed?

Alex McMullen:
Believe it or not, the conversations around CMDBs and ITL were actually a formative part of these discussions. Really what Fusion is delivering from a consumer perspective is two things. It’s storage classes and protection policies. And when you marry those together in whichever combination you wish, you then are able to define a series of use cases that are actually relevant for every business, not just traditional things like production and development, but also different capabilities by region or by cost profile. Maybe it’s cheaper in Azure than it is on-premises today so I’ll have all my backups, snapshots being redirected there, for example. But really this is very much that policy based method of consumption of the future.

Alex McMullen:
The persona that’s actually funding and driving utilization of devices in this use case, isn’t interested in infrastructure. They’re interested in the capabilities that it can bring to help them in their day job. And that’s really where we see Fusion being so powerful because it allows the storage and the admin and the data center personas to define the capabilities of the classes the stores are willing to offer in every region, in every availability zone. And then they also in conjunction with whether it’s your internal compliance team or their legal team, they define protection policies that are also available. And then by marrying the two of those in the different classes of storage, Fusion allows customers to pick and choose the ones that suit their business requirement. And that I think is a very clear, simple, but also hugely powerful message.

Chris Evans:
Absolutely, I agree. Now, imagine I’m a customer, I come along to you, and I’m a customer who, as of today, when you’re doing this launch, I’ve maybe bought a few products from you and I’ve got them in the data center. And maybe I’ve started to look at the idea of maybe taking things as a service and not actually buying the hardware. How am I going to go forward with the idea of Fusion with this mix of different technologies that some might be owned, if you’d want to use that term, some might be actually just a consumption based model. How’s this going to all fit in to allow me to, A, balance all of that out and make sure I’m not under-using my own resources, but I guess work towards that transition that says, eventually at some point I want to get to a fully service model. How’s that going to work with Fusion from today?

Alex McMullen:
I guess if you’re a Pure customer and thank you for your business, obviously, you’ll already be familiar with things like non-disruptive updates and hopefully you’ve done an Evergreen cycle because once you’ve done an Evergreen upgrade, that’s when customers really start to get why Pure is different. It’s that my RA is three years old. Oh, hang on. Now, it’s really new again and nobody noticed. So that perpetual or long-term depreciation cycle of 10, 15 years really has changed that game.

Alex McMullen:
But the real conversion theory we’re talking about here is that the power of combination of Evergreen and Fusion is that, if a customer wants to run future purchases as a utility, that’s fine. If you want to run existing hardware as is without further charge, then that’s also fine. And in further waves of the product, will actually allow you to have part of the Array or series of Arrays in the Fusion persona and some of those in the traditional personas. So the Fusion software only manage proportions of the Arrays, and that’s how we’ll retrofit it into existing deployments where let’s say the Array is 90% full, then maybe Fusion will manage 5% until you want to migrate things away or move them around to make more space.

Alex McMullen:
So the great thing about Pure software flexibility is, we’re able to cope with a whole range of the different commercial challenges around. And quite honestly, if a customer wants to with their existing fleet, convert all of that into utility-based running and purchasing, we have a financial structure to do that too. So there’s no lost investment. Pure is very proud of its long-term and longstanding customers and we’ve always been very focused on maintaining the same level of capability with them as we have with new customers. And those terribly old cliched adverts of brand new customers only. That’s very much the antithesis of Pure’s view of the world. So people with the deployments they have will be able to have partly or all in Fusion if they wish, and then go forward with more traditional CapEx purchases or on a utility basis as well. It’s completely up to them.

Chris Evans:
Brilliant. I really wanted to just emphasize that because I think in the way that we’re moving forward, it’s very easy for people to just look at the technology view of this. But in reality, we’re moving further away from the technology view. Although the technology’s important, it’s becoming more hidden, shall we say, more behind the scenes, more going to potentially be managed by companies like yourselves.

Chris Evans:
So what’s going to be more upfront is going to be the business relationship, is going to be the consumption model, is the cost management from the business and the customer to understand how they actually can best use the resources available but also come to you and say, “Well, what’s the best route forward as I evolve the fleet of technology I’ve got?” So how will customers be charged for using Fusion? Will it become part of their existing Pure1 that they get for free? Or is this going to be additionally charged? How do they optimize what they can be paying?

Alex McMullen:
We’ll finalize the pricing structure once we complete the early access program. In the early days, that will be something we’re actually most interested in talking to customers about. Are they interested in doing it on a per device basis, on a per gig per month basis or some combination of the two? I think what I will say right now, I’ll be confident in stating that whichever way we go with it, customers and existing Pure customers will be happy about that. It’s really one of the things that we find most interesting about the discussion is, how do customers want to be charged for this capability? We’re happy to do either at this point in time.

Chris Evans:
Okay. And we’ll wait to see what that turns out to be then. I know that we’re talking today and today’s the announcement point for all of these products, but there’ll be an early access program and eventually there’ll be a general availability at some point. Early access is where we’re going to be starting first, isn’t it? For a few months I see.

Alex McMullen:
Yes. This is a very sophisticated piece of software with lots of different components. Some containerized, some web front-end, some on-prem. So it is one of the things we want to get completely right and because obviously with the security aspects in terms of control from the public cloud within the perimeter of our customer base, we want to make sure that link is also 100% perfect. So it’s a very structured EAP, early access program that we’ll be starting. In fact, it’ll be in two waves. The first wave is more likely to be the Portworx Data Services side, which we’ll talk about in a moment, but we anticipate both the programs will complete before the end of the year. And then we will launch them early next year.

Chris Evans:
Great. I just want touch on one little thing before we dive into PDS, if you want to call it. Partition data set, which is all I ever see when I see PDS. That shows how old I am. And that’s really to just talk about the backend work that you obviously have had to put in to make this work. It’d be good to understand where Fusion will be running, because I think people even though it’s a SaaS offering, will have reservations about understanding where that was sitting. But secondly, just to understand how much additional work you have to now have going on in the background with these new services, in order to be able to make sure you can run them 24 hours a day, be responsive to customers and so on and in a global environment where people could have equipment that’s in every time zone around the world. So what’s changed within Pure to get you to that point?

Alex McMullen:
I think the biggest, I guess, change of the way we look at the world has been in terms of the transition into effectively an SRE mindset for not just the dev team, but also the support teams. Because obviously there’s lots of new concepts that they have to learn as part of this whole Fusion layer. So our focus has been primarily on the machine learning, the analytics side of things, the telemetry, and that transition from Pure1 being primarily a monitoring aspect to now being not just observability and predicting the future, but also in terms of active control of devices over a network.

Alex McMullen:
So the way we’ve looked at this is, we already know the way that the Array behaves in phones home telemetry and things like that. So that’s all taken care of but the aggregation of that into what’s happening within an availability zone, or in fact, even a region for those layers, as well as seeing the plan pipeline of data movement that’s being intended by Fusion as part of this workload balancing or new workload deployments, all those factors have made us focus a lot more on the observability side of the house. And that’s really what’s being, for me most interesting is watching that change in shape for engineering, moving to faster merges, moving to that focus on security of links.

Alex McMullen:
The key thing for it is obviously, whilst the application, the SaaS portal layers run in public cloud, it needs to be able to maintain those controls. So we have a focus on making sure the application remains available. If we did lose a number of cloud regions, then obviously customers won’t be able to make changes. They’ll be able to see what’s happening and the current connectivity will continue to remain, but until the cloud then comes back, they’d obviously be without the ability to make any more changes. So we have to factor those aspects in.

Alex McMullen:
It’s become a much more worldwide challenge I think Chris, in summary. It’s very much more a global 24/7 picture. We’re now focusing on much more of a life lesson in terms of how containerized apps work in multi-region basis and of course the security aspects that go with that given the way the world is focused over the last 18 months, I guess, in terms of supply chain and ransomware and zero day disclosures. So there are many more things for us to think about but that’s just a quick broad sense of some of the things we’ve now taken on board as part of it.

Chris Evans:
Great. I think this is a nice little segue actually in terms of a transition there. But I think when you talk about that ability to be able to make changes or do things to the infrastructure, I would suggest that as you head forward, those sort of changes are not necessarily going to be happening at that physical hardware level very often. And part of the reason for that is that you’ve started to obstruct up certain features into things like Portworx and the Portworx platform. So rather than having a high churn on things like creating new volumes and LUNs and restructuring the data, a lot of those services are being pushed up to a slightly higher layer and we start talking about that next layer up in our layer cake, or maybe a few layers up where actually the infrastructure might be more stable, more static, but actually other pieces start doing that work. And this leads us into the discussion about Portworx and Portworx Data Services. So I would hope that most people know that you acquired Portworx last year, or the year before. I can’t remember the exact date.

Alex McMullen:
Last year.

Chris Evans:
Last year. Gosh, I forgot where we are in terms of [inaudible 00:28:19].

Alex McMullen:
Yeah, you lose a year here and there, don’t you?

Chris Evans:
You’re not kidding. Yeah, you certainly do. Obviously you acquired Portworx, Portworx is a container attached solution or a obstruction layer at the contender layer for storage. What is PDS?

Alex McMullen:
PDS is our second SaaS product announcement of the day, and really it’s that SaaS layer to deploy database as a service within containerized applications, which are managed by Kubernetes. So think about the next level of challenges. As you said Chris, Portworx itself is software-defined storage for containers, and it already has a number of very sophisticated capabilities. Some of which, frankly, we’re bringing into Fusion as a reverse segue. So it knows today how to do workload balancing, how can we do data movement, how can we do again, predicting for the future.

Alex McMullen:
And some of that is being driven of course, by the Kubernetes cluster itself. But really what Portworx Data Services here is bringing to the container world is the ability to deploy a range of modern SQL and no SQL databases automatically into a Kubernetes cluster, but also much more importantly, it’s able to fulfill the day two operations that go with that in terms of high availability, backup, restore, protection and a whole range of other things you’ll see as we go forward. So very simply Portworx Data Services database is a savers layer for Kubernetes clusters.

Chris Evans:
Now, I’m really interested in this and the reason I say that is because we’ve been talking about container attached or the software defined storage layer that sits above our hardware for a while. When we first started talking about using things like CSI to administer storage Arrays, it did make me think, well, this isn’t really likely to work very well because the first thing you have a problem with there is that usually, in this type of environment, there’s a higher churn on creating and using volumes. And there would be at the physical layer and the storage platform and most storage platforms just aren’t designed for say, hundreds of LUN creations an hour or even thousands of LUN creations an hour.

Chris Evans:
The Portworx solution has that nice abstraction layer, but it obviously is a lot more than that. And as a result, you now got a situation where you’re building that next layer up, but you’re able to add in those additional features that actually are more application specific. And now you’re deploying the actual applications at the same time.

Alex McMullen:
Yes, very much so. And it’s again an echo of feedback from the customer base, as well as watching how the technology world is changing. One of the great lessons that we saw pretty early on after the Portworx acquisition last September was really the view from some of their big customer deployments. As you said Chris, the rate of change of objects within high-performance Kurbernetes applications, whether that was Royal Bank of Canada or roadblocks, or some of the airlines, it really brought home to us. It was that really you’re adding and deleting thousands of these guys a second. We don’t have a story for that. So it was very much accelerating our own decision to bring Portworx as that layer and have that Portworx capability deployed as standard within Pure devices and that absorbed our CSI functionality. And that’s the way we go forward, because it is very much about embracing the modern requirement and that’s how we got to this point.

Chris Evans:
So exactly what will people be able to do with PDS? Are they going to just be able to pick a database from a list, deploy it and have it just be spun up? What’s the user experience going to look like?

Alex McMullen:
Very simply, Portworx Data Services again, as a SaaS offering, you had a portal which is hosted in the public cloud, which has all the usual Portworx support credentials that allows you to log in. That gives you the availability to register one or more of your communities clusters with the Portworx API. And then by doing so, that allows you to first obviously install the Portworx data service operator, which is where most of the brains lies here. And that will then allow you as that customer to deploy from a curated list of database applications, which are tailored toward a different range of use cases in that containerized world. So we’ll do some for streaming, some productive management, some for time series, some for relational work. And really as a customer, you say, “I want that database, I want it to be called this and I want it to be this capacity.” And then Portworx Data Services does the rest.

Alex McMullen:
What you get back from that is, essentially a connection string in the API, which allows you then inside your app to go and add, modify, delete data in all of those databases. The most important part of this really is though, it’s not so much the deployment because that’s something you can do on a manual basis. But what it does do is, deploy repeatably but also then using that operator technology that we’ve written is the day two side of things. So using the Amazon day zero, day one, day two analogy. The operator itself will actually monitor, maintain and protect the database that you’ve just deployed, which is unique in the industry. I believe that’s an industry first as well.

Alex McMullen:
So it takes care of all of the issues in terms of, oh, it seems to be broken, all right. If it’s a shorter database, do we have the anti-affinity to make sure it’s all running on the same nose. And then backup recovery, it’s one of the biggest challenges in containerized applications today is data protection and that’s all taken care of by PX-Backup, which we’ll come back to also. So it’s not just the deployment, it’s the day two capability. The ongoing again Prometheus metrics, again, with PX-Backup, again with the ongoing, making sure the lights are all still on.

Chris Evans:
This is a really interesting part of our industry I think in that, every time we come up with something new and we use the virtualization example is a good one. So virtualization and then originally Docker, and then Kubernetes as part of this container evolution, if you like. As we look at all of these things, the first thing we say is that they come to the market and people go, “Brilliant. Wow, that’s amazing. That’s great technology.” And then you’ve got to start thinking about, well, actually operationally, how does it fit into the way that I run my business? So I can’t do something unless I can guarantee I can protect the data, I can potentially encrypt the data, I can manage it so that I’ve got high availability and all of these things have a habit of coming along a little bit later on.

Chris Evans:
And I suspect that we’re now in that period where we’re starting to see the adoption of containerized applications, getting to the point where people are saying, “How do I deal with these pieces of the infrastructure? How do I manage the,” as you call them , “The day two pieces?” And certainly I would suspect that most developers really couldn’t care less about having to worry about that if they could avoid it and being able to push that off to a component that actually delivers that as a service means, they can be more focused on what the application does and less focused on managing the infrastructure, hence the reason why we are where we are.

Alex McMullen:
Yes, very much so. And the whole driver for Portworx Data Services going forward as to widen that list of databases that we will support, deploy, manage. The key thing about it is really that one of the really great things about having container registries and having that curated list of images that you can pull down, is it all nicely secure, nicely access controlled, but it’s really that ongoing focus on future challenges, modern problems, today’s developer issue, tomorrow’s data scientist issue. As you say, they just want to deploy whether it’s a Kafka Stream or a Mongo instance to suit their application in that 12 factor mindset of having the database as something over there that’s perhaps not relevant for my application, but I need it, I need to use it.

Alex McMullen:
So the concept of bringing back, for example, the connection string and putting that into the CRD, letting them consume those credentials via the standard community’s APIs. We’re giving them the things in the way they want to be used rather than how the use has evolved over the last 20 years. Again, taking that big step forward in simplification and that standardize, yep, I can use this to do my day job approach rather than the complexity that sits underneath.

Chris Evans:
And to a certain degree, I think you’re effectively moving up the stack, but at the same time, following to a certain degree, the same concepts as you would do if you were consuming physical resources. So if you want to consume physical resources, now you can do that via an API, you can, in your instance, use Portworx to do that and say, “Give me a volume as part of the application starting up.” Now, we’re saying, “Well, I don’t just want the volume. I want a complex piece of storage. I want a piece of storage that actually understands how to manage data to a different type of protocol level.” And this instance, a database protocol. So what you’re actually giving me is, you’re giving me another object for storing data, which is a bit more complex, but actually it’s still requested, managed and delivered to you via an API and then you just consume it from that point onwards.

Alex McMullen:
Yes, very much so. I don’t want to get ahead of myself here, but some of the things, the problems we’re hearing from developers today in terms of, how do I clone that database that I’m working on? How do I wind it back to yesterday? Again, that [inaudible 00:37:11], I guess modernization of CI pipelining, those are the sort of things we’re looking at now, how do you manipulate a stateful set into something that can be used in exactly that way? Those are all the problems we’re looking at next.

Alex McMullen:
But the focus for today is on bringing together that ruthless simplicity that Pure is famous for, but allowing incredibly capable product like Portworx, which has a number of enterprise features already and it can do data move and data migration, and it can do snapshots already to bring that up a layer to solve database problems and consumer problems as well of course. That’s really the big focus. And that’s why the abstracting of all this into that SaaS layer, that top layer on the layer cake, if you will, that’s what really brings it all home and smooths over. Well, let’s be honest. There are still some cracks underneath in terms of Kurbernetes manageability and simplicity and security as well. But what we’re doing here is, smoothing out the road for the people trying to consume it as best we can.

Chris Evans:
I’d like to just ask the question about what managed actually means because manage could be as wide and deep as you’d like to think of it. So it could be anything from in-place upgrades to, as you said, correcting corruption issues or correcting resiliency issues, all sorts of different things. So in your first instance, what will managed really cover and it’s like then they’re unlikely to expand over time.

Alex McMullen:
Yes. We’re broadly following the same guidance as for example, Amazon does with Redshift and with various other database products it has. So what’s the Portworx Data Services operates. So it’s an operator that does this in terms of their container terminology, it essentially collects different metrics depending on which database it’s actually managing and monitoring at that point in time. They get phoned home to the SaaS layer and if there was something in the logs that we don’t like, or that we think is an issue, then the operator will take specific steps to correct whatever it thinks that particular aspect is.

Alex McMullen:
But just to be clear, if we encounter a bug, for example, I’m just picking one now, if I encounter a [inaudible 00:39:13] bug, you as a customer, still need to resolve that in the usual way because that’s not something we can deal with. We can only deal with runtime issues that you would encounter on a day-to-day basis. So it’s looking at the health state of database. It will do auto expansion on the capacity side, it will solve for the usual in terms of the node has gone, the application has failed. Those will get restarted, re-corrected. Shards will get moved as required if it’s sharded, but for software issues within the product itself, you go via the usual route.

Chris Evans:
I’d like to put a comparison in here because you mentioned earlier when you talked about RAID and I had to bring in the mainframe, but I’m going to bring in the mainframe as an example.

Alex McMullen:
Wow, okay.

Chris Evans:
You look back to say, when I first started working in the mainframe world, we very much physically, as in storage administration, would be responsible for making sure that data was recovered because there was no red built into that infrastructure. And it was very hands-on, very specific and you couldn’t really manage that much in terms of capacity. The introduction of things like RAID took away huge amounts of the tedious side of work, of having to recover data. You trust that the Array will do that for you. And over time, I would suggest that a lot of the more, not menial, but certainly the more repeatable tasks that could be automated have been phased out or dealt within in the hardware and the software.

Chris Evans:
And this seems to me that what you’re offering here is the same thing too, because you can imagine DBA is thinking, “Oh my God, I’m out of a job.” Because there’s now all of this automated deployment and management. Well, actually in reality, it’s more likely to be that you’re taking away those more basic tasks of deployment and management and upgrades and fixing that allows the DBA to go and focus more on the data and those more high level functions. So you’re not doing anybody out of a job here. You’re just actually doing exactly what we’d like to see happen in the industry in general.

Alex McMullen:
Yes. So having worked at a number of investment banks in the last couple of decades, I can tell you for sure that the jobs that DBAs hate most are database refreshes, database restores, new build deployments and getting storage allocations and building all that layer up. They want to do database level tasks. They don’t want to be involved in all that infrastructure stuff. And many of the investment banks spend a lot of time and investment engineering-wise to give utilities that would do all of those things that they could then go and run.

Alex McMullen:
So this is very much not about making DBAs less useful. It’s simply allowing them to focus on the things that actually bring a business benefit. And honestly, if I could have offered the DBAs 10 years ago in response to a help desk ticket, click this link, there’s your Oracle database out of being the hero at that point in time. So this is very much bringing that level of capability into that containerized Kubernetes world and it’s really going beyond that. People in today’s personas don’t want to have a ticket into service now or into some remedy database, they want to click and go. And that’s what really is most powerful about Portworx Data Services here.

Chris Evans:
Perfect. And that’s exactly what I would agree with the way I see it too. So how does this now integrate into Fusion Pure1? Can I now see all of my managed databases within that same infrastructure? And will I get a holistic view of all this?

Alex McMullen:
Yes, you will. In terms of the integration points, all of this is being driven by Pure1. Pure1 is that go-to portal for SaaS layers for utility billing. It’ll be for management on the Fusion side as well. So you will expect that single picture. In fact, there will be screenshots in some of the material that goes out today that show communities clusters being run and this structure being shown within Pure1 as we do today already with VMware and the VM analytics side of things. So you’ll get exactly that same picture being brought back to you. And of course it’ll be Grafana dashboards, all the good things that come via Prometheus as well. Is that very much modern interface to the modern application and we’re not trying to go back to the days of a VT320 and green text on a black background. This is all modern, elegant, as you’d expect in that cloud-like persona that we very much have here.

Alex McMullen:
In regards to how Fusion works with Portworx Data Services the thing that we want to make sure we have most is complete integration and not getting ahead of the CSI spec on the container side. So we do want to add functionality that will then either be different in CSI or not in CSI as well. So the cadence at which the Portworx Data Services extra functionality with Fusion will be adopted, will depend on the rate of growth of that spec. You could certainly be 100% confident that you’d be able to run all kinds of things, whether it’s vVols on Fusion with Kubevirt with some other weird obstruction layers as well, they will work perfectly together. Yes.

Chris Evans:
Great. I think this is a really interesting evolution of the technology. And just from what we’ve said already, I think it’s going to be fascinating to see. What I’d like to just, I guess, finish on in this part of the discussion is to understand exactly where this will be deployable because obviously in terms of Fusion, that’s a management of Pure Storage products fairly clearly. Portworx to a certain degree is very much independent if you want it to be from the hardware platforms. So will I only be able to use PDS to deploy infrastructure that’s sitting in the public cloud as well as on-premises? Is that the reason why you say that you’re trying to keep within the CSI spec so that you don’t break anything for customers who might be looking at the public cloud for that deployment?

Alex McMullen:
Very much so. The Portworx ethos and our strap line there is any, any, any. Any application, anywhere, any cloud and Portworx Data Services aims to echo on that. In fact, it builds on top of Portworx enterprise. So anywhere that Portworx enterprise will run, you’ll be able to run Portworx Data Services. Whether that’s AWS as your Google Cloud in hyperscaler or colo data center and of course on-premises as well. So Portworx is software defined storage, anywhere it runs, you’ll be able to deploy on Portworx Data Services. The only thing you require is the ability to permit Portworx Data Services to access your Kubernetes clusters. And that’s the only correlation required.

Chris Evans:
Perfect. And dare I ask about pricing and early access programs in the same way that we talked about Fusion because I suppose we have to ask that question, don’t we?

Alex McMullen:
Yes, we do. So the early access program for Portworx Data Services starts today. The first day is today. So anybody who has an interest in deploying databases as a service within a Kubernetes cluster then, click on the links that are at the end of the show, or obviously watch the webcast that are going out live today as well. The big takeaway here is, it’s really, really about the application for the future. And it’s really down to… The only limitation is with you as customers in terms of how ambitious you want to be within your own infrastructures.

Chris Evans:
Alex, it’s been really interesting to understand the two main features you’re delivering here and this appears to be part of what I can only, I think you have used this term digital experience. The way the company’s transforming to give you much more through software and make it easy to consume but you’re not give up on hardware, are you? I mean, what does this say about the direction of where Pure is headed and how the strategy in a play out for the next 10 years, what can we expect to see?

Alex McMullen:
No, we’re not going to stop doing hardware anytime soon. But let me just summarize then I guess what we’re talking about here. Pure, its heritage has been in terms of its software strategy. We took the business decision in, I guess, 2013, ’14 to make our own hardware platform. And we’ve done that for reasons that are very obvious now for our customer base in terms of that perpetual blade chassis capability and that modern refresh every three years, which is so incredibly powerful. We remain fully committed to hardware and we make our own drives, we make our own compute blades. So that will continue.

Alex McMullen:
Our platform BU, is world-class and we believe our hardware solutions will continue to embrace all the modern, great things, not just in terms of the flash side, but also developments and PCIE buses, the CPU roadmaps, and some of the other DPU, other abstract capabilities that we want to embrace where it makes sense. So absolutely we are still a hardware company. The reason that we’ve been so successful is the ability to evolve our software capabilities on top of that. And really what you’re seeing here today is the first real site of Pure with its cloud ambitions, with its software as a service ambitions, doing that across all of its hardware fleet, doing it in the public cloud as well, and then doing it for containerized applications for Kubernetes as well. That’s an incredibly powerful message for those folks who had us originally tagged as, oh, they must be people that make boxes.

Alex McMullen:
We do make boxes, they’re the best boxes out there, but also they have world class state of the art software, not just on the platform, but you’re seeing that bridging layer across that, embracing the public cloud and embracing those cloud methodologies and capabilities in the same way they’ve always done, with backward compatibility, with commitments to the other parts of the industry and remaining software defined with Portworx. So pretty broad ambition for us. And we could talk about some of the long-term strategy things too of course.

Chris Evans:
I’m glad to hear you’re not giving up on the hardware and I didn’t think you ever would but that’s been a real piece of the story that we’ve talked about many, many times before and some of those features that are in there that are really relevant to how you now deliver the software. And I think the message for me is, I think, looking at this is that the two things need to be fairly closely coupled. You can’t just say, “Well, I’ll just take the products I was selling to the customer on-prem last week.” And now say that they’re all cloudified and you can buy them as an on demand and think that’s going to be an efficient way to deliver technology. It just doesn’t work like that.

Chris Evans:
We know there’ve been lots of different things in the technology that’s got us to that point. And I will, at the end of this, I’ll link to some of the other pieces where I’ve discussed that in the past and we’ve talked about it. I guess, we’re seeing an evolving company here, but a company that still has a strategy that’s been there for a while, as you said, 2013, 2014. And that perhaps reflects what we’ve said at the very beginning of this discussion that customers are looking for something different in terms of their experience. This is all about providing what the customer wants because the customer is adapting to using technology in a different way.

Alex McMullen:
Yes, they very much are. And it’s also much more societally influenced as well. If you think about what’s happening in the wider world, one of the reasons that Pure’s message is so powerful is because of its sustainability position, not just in terms of saving data center power and cooling because we’re small, you only need to 1,000 watts to run your storage Array rather than 15 or 20 as it used to be for the other folks. But it’s also about the supply chain side in terms of reusability and keeping that same storage Array and drives for 10 years, if you wish.

Alex McMullen:
So cutting down on waste, limiting recycling and that minimizing our carbon footprints is a powerful and strong message from many modern CEOs who are asking, “I’m building my own company for the future. What are you as my partner, Pure Storage, doing about commitments to the environment and to sustainability and carbon sequestration, et cetera, et cetera?” And we’re working very hard on all those spaces. So having that story, not just a veneer, but an actuality has been a very compelling part of our company vision, our company position for the last few years. And you’ll see us increasingly focus on it also.

Alex McMullen:
So keeping it modern, but also keeping it in a way that’s sustainable for, the wider planet as we continue to change shape over the next few years. We will do more software, we will do more hardware platforms, but they will work backwards and forwards. They will work beautifully together and we’ll continue to operate with our key technology partners, not just in the cloud, but also for data protection as well, for example. So that’s really why we think it’s keeping itself modern and relevant for today’s consumer.

Chris Evans:
Brilliant. It’s been a really interesting discussion. By the way, on that whole reusability and carbon footprint and all the rest of it, we did actually think about doing a whole podcast series talking about that because there’s a whole raft of things that can be talked about in that area. And when you look at last week and we’re recording this in the UK, you look at the issues that we saw last week around fuel and the gas supply and costs going up, the cost of actually deploying and running infrastructure is now becoming a really significant thing to consider.

Chris Evans:
So all of those messages I think, resound and are very important in what we’ve just discussed. So really interesting day-to-day to see all this stuff come out. If people want to find out a bit more, we’ll put some links in the show notes, but where should we point them to from your perspective at this point?

Alex McMullen:
Well, I think the easiest thing quite honestly with Chris is knowing how marketing teams work. I’m just going to say, look at the Pure website. I’m not going to try and spell URLs because I’ll get them wrong. So start with purestorage.com. I’m very confident it’ll be there in big splashes, details, messaging, pricing, structure intent, as well as the broader capabilities of both Fusion and Portworx Data Services.

Chris Evans:
Perfect. Alex, thank you for joining me today. It’s been a really great discussion. Congratulations on this launch. And it’s really interesting to see the evolution of the company. And guess what? We’ll catch up in six, 12 months time and talk about this all again and see how it’s gone.

Alex McMullen:
It’s been a pleasure Chris. Thanks for the time.

Speaker 1:
You’ve been listening to Storage Unpacked. For show notes and more subscribe at storageunpacked.com. Follow us on Twitter at StorageUnpacked or join our LinkedIn group by searching for Storage Unpacked Podcast. You can find us on all good podcast catches, including Apple Podcasts, Google Podcasts and Spotify. Thanks for listening.

Related Podcasts & Blogs * Architecting IT Pure Storage Microsite * #209 – Discovering Unified Fast File and Object with Pure Storage * #199 – Quantifying Data Storage Innovation * #198 – Software-only Storage Vendors * #185 – Pure-as-a-Service 2.0

Alex’s Bio

Alex began his industry career after Bachelors and Masters Engineering degrees at Liverpool University. He started by writing flight/test software at British Aerospace, CAD 3D modelling software at Parametric Technology and the RDBMS engine at CA for Ingres before moving into infrastructure design and implementation. Alex transitioned into professional services at Sun and VERITAS through a consulting partner with numerous assignments designing and implementing large scale UNIX and storage systems in financial services. Alex then joined UBS Investment Bank to manage the storage and distributed computing infrastructure for the bank, as well as assessing innovative and emerging technology in Silicon Valley. From UBS he moved to Barclays to run the storage, database and data warehouse engineering divisions and spent 6 years there before joining the Pure Storage team in 2014 where he holds the role of VP and CTO, International.

The post #211 – Pure//Launch – Announcing Fusion and Portworx Data Services (Sponsored) appeared first on Storage Unpacked.

View Details

This week, Chris and Martin discuss the merits of building virtual SANs in the public cloud. Vendors including Silk and Pure Storage now offer virtual storage “appliances” built from virtual instances and cloud storage. Why are these solutions necessary, when the public cloud providers have plenty of high-performance block and file storage offerings? The discussion looks at the why and how of these types of solutions as well as the implications and the lock-in they could represent.

Elapsed Time: 00:33:53

Timeline * 00:00:00 – Intros * 00:04:00 – Silk and Pure Storage are two examples of cloud SANs * 00:05:55 – Why build a SAN with public cloud infrastructure? * 00:08:00 – Are affinity rules available for public cloud compute? * 00:09:45 – Cloud vendors used to recommend striping LUNs for performance * 00:11:00 – Are these solutions about fixing latency problems? * 00:12:10 – Two models – dedicated storage or HCI * 00:14:27 – Could the benefit for Cloud SANs be functionality? * 00:16:10 – We’ve gone back to 1999 – SAN Volume Controller * 00:22:50 – Could the CAS platforms have an advantage for SAN storage? * 00:25:00 – Pure Storage is already integrating Portworx into array automation * 00:27:20 – How do users calculate cost savings of Cloud SANs? * 00:32:00 – Cloud SANs could be all about saving on-premises costs * 00:32:30 – Wrap Up

Related Podcasts & Blogs * #207 – AWS Gets a SAN Upgrade * #196 – Creating a Multi-Cloud Data Strategy * #179 – The Myth of Cheap Cloud Storage * #116 – Fixing Gaps in Cloud Storage with Andy Watson * Pure Storage Cloud Block Store Goes GA * Cloud Services – Build, Buy or Fork?

The post #210 – Building SANs in the Cloud appeared first on Storage Unpacked.

View Details

In this week’s episode, Chris and Martin meet with Brian Carpenter, Senior Director of Unstructured Technology Strategy at Pure Storage, to talk about UFFO – Unified Fast File and Object. UFFO is a technology bringing together the two main types of unstructured data into a single platform – FlashBlade. The discussion centres on the benefits of using a single platform that offers high performance throughput for both file and object at the same time. Brian walks the team through use cases and how customers are using UFFO in their environments today.

You can find out more about UFFO at the Pure Storage Website – https://www.purestorage.com/products/file-and-object/flashblade/unified-fast-file-object.html which has details on the use cases and technology. Learn more about FlashBlade here – https://www.purestorage.com/products/file-and-object/flashblade.html

Finally, here’s a link to the Tech Field Day presentation by Brian Gold. https://techfieldday.com/appearance/pure-storage-presents-at-tech-field-day-23/

Elapsed Time: 00:42:20

Timeline * 00:00:00 – Intros * 00:01:00 – UFFO – Unified Fast File and Object * 00:02:00 – Why do we need fast object & file? * 00:03:40 – Object storage has been evolving for years * 00:04:35 – Object storage doesn’t mean petabytes of capacity * 00:06:05 – FlashBlade has been in the market just over 4 years * 00:07:45 – All-flash for backup? It depends on RTO/RPO * 00:10:35 – File/Object performance is critical with modern media * 00:12:25 – UFFO enables customers to refactor their data without rearchitecting * 00:14:40 – Applications now mix file and object usage * 00:17:35 – FlashBlade isn’t mixing protocols on the same data * 00:20:20 – FlashBlade enables growth per blade, with persistent performance * 00:24:06 – The FlashBlade design makes it easy to build scale-out systems * 00:26:25 – Use cases for workloads grow incrementally once systems are installed * 00:29:55 – Martin still believes in a separate silo for data protection * 00:32:05 – Cheaper flash will see greater adoption of UFFO in the data centre * 00:34:45 – What are Pure Storage customers using UFFO to achieve? * 00:37:55 – Amdahl’s Law is still relevant today * 00:40:10 – Wrap Up

Related Podcasts & Blogs * Soundbytes #008: FlashBlade 2.0 with Rob Lee at Pure Accelerate * #118 – Pure Accelerate 2019 with Patrick Smith * Pure //Accelerate 2016 – FlashBlade * Moving to Unstructured Data Stores

Brian’s Bio

Brian Carpenter is an experienced chief technologist, public speaker and thought leader. As Senior Director of Unstructured Technology Strategy at Pure Storage, Brian enables customers’ journeys to high-performance outcomes with Artificial Intelligence, Advanced Analytics, and Platform as a Service. Brian’s served in a variety of roles including heading up IT Departments, creating and fostering technology alliances, and business development initiatives for the data storage industry. He is a trusted advisor with real-world customer experience who can translate complex issues (and common struggles) to both technologists and executives alike. Brian shares his passion for technology as a business enabler through mentoring programs and advising startups on IT strategy.

The post #209 – Discovering Unified Fast File and Object with Pure Storage (Sponsored) appeared first on Storage Unpacked.

View Details

In this week’s podcast, Chris and Martin look at NVIDIA’s GPUDirect Storage (GDS), a technology for moving data directly from persistent storage to GPUs, bypassing the CPU. The aim of the technology is to provide greater throughput to keep GPUs active, and looking at some of the thoughts from our podcast with Liqid (#204), it’s clear that this feature is needed. However, with such quick adoption by many storage vendors, is the announcement of GDS more of a marketing exercise? We dig into the details and what the announcement of GDS could mean for the future enterprise.

On a side note, there was indeed a remake of Magnum P.I. in 2018, following on from the original series in 1980. The original was a classic; we’re not sure about the remake.

Elapsed Time: 00:30:42

Timeline * 00:00:00 – Intros * 00:01:00 – The Chiacoin price has collapsed! * 00:02:15 – What is GPUDirect?00:06:05 – GDS has Local and Remote modes * 00:07:15 – GDS appears to be a data mover offload * 00:09:05 – High throughput application look to benefit most from HDS * 00:10:00 – GDS is not a generic replacement for external storage I/O * 00:11:40 – Many vendors supported the announcement of GDS * 00:12:53 – What drivers and mods will be need for GDS? * 00:14:38 – Disaggregated architecture strikes again with ConnectX * 00:16:04 – GPUs are a next-generation data centre technology * 00:17:43 – CPU lock-in is a thing – could GPU lock-in exist too? * 00:19:28 – Will NVIDIA lock Arm and GPU technologies together? * 00:23:29 – Is the Von Neumann architecture dead? * 00:25:30 – Will all vendors eventually support GDS? * 00:30:00 – Wrap Up

Related Podcasts & Blogs * #204 – Liqid Composable Disaggregated Infrastructure * #190 – NVIDIA BlueField SmartNICs and DPUs * Intel Under Pressure as NVIDIA Announces Grace CPU * Persistent Memory in the Data Centre

The post #208 – NVIDIA GPUDirect Storage – More Than a Good Marketing Message? appeared first on Storage Unpacked.

View Details

This week the team is back after a short break and ready to talk all things cloud – or at least about Elastic Block Store. AWS recently announced io2 and io2 Block Express, the first new EBS block storage offerings for eight years. With such a long gap since io1 was first released, what has AWS been up to and what’s new? In this discussion, Martin and Chris question whether AWS is aiming to meet the requirements of enterprise customers that simply want to lift and shift their infrastructure into the cloud. Alternatively, the effort involved in rolling out a new distributed storage platform is significant, so this could be all about time and effort.

Elapsed Time: 00:41:28

Timeline * 00:00:00 – Intros * 00:02:15 – Is storage done or has there been a COVID slowdown? * 00:03:05 – Is AWS running a secret SAN? * 00:04:30 – What is EBS – Elastic Block Store? * 00:07:30 – AWS finally offering five 9’s on EBS * 00:08:20 – io2 and io2 Block Express add new availability & performance * 00:09:40 – Latency is a hard storage attribute to manage * 00:13:00 – How has EC2 evolved in capability? * 00:14:20 – Was AWS caught out with block storage demands? * 00:16:25 – Nitro is a big driver in new io2 storage (we think) * 00:19:00 – How would AWS refresh the storage infrastructure? * 00:21:45 – Could you “triage out” slow EBS volumes? * 00:23:25 – What if a DR location runs slower than primary? * 00:27:00 – Does Nitro indicate the direction for enterprise storage? * 00:31:00 – There’s no (customer) synchronous replication in cloud * 00:32:00 – What comes after EC2/EBS? * 00:34:50 – Protocol wars could be the next battleground for storage! * 00:35:40 – Does refresh (and storage refresh) matter in the cloud? * 00:39:00 – Wrap Up

Related Podcasts & Blogs * Disaggregated Storage Part II with Zivan Ori from E8 Storage * AWS is not building Hybrid Cloud (as we know it) * AWS Outposts – A Cuckoo in the Enterprise Nest?

The post #207 – AWS Gets a SAN Upgrade appeared first on Storage Unpacked.

Storage Unpacked Podcast: Recent Episodes

Storage Unpacked Podcast

234 – Introducing Pure Storage FlashBlade//S and Evergreen//Flex with CTO Rob Lee

232 – Building Power-Efficient Storage Systems with Justin Emerson from Pure Storage

234 – Introducing Pure Storage FlashBlade//S and Evergreen//Flex with CTO Rob Lee

232 – Building Power-Efficient Storage Systems with Justin Emerson from Pure Storage

209 – Discovering Unified Fast File and Object With Pure Storage

133 – FlashArray//C Deeper Dive with Pete Kirkpatrick