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

Copyright (c) 2016-2021 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #h17t.

View Details

This week's podcast episode continues the discussion on SmartNICs and DPUs with Fungible, a company that claims to have originally coined the term DPU. Chris and Martin talk with Pradeep Sindhu (CEO and co-founder) and Jai Menon (Chief Scientist) about Fungible's storage cluster and host-based DPU. The Fungible architecture aims to solve the challenges of disaggregation, a topic we first looked at back in September 2017.

This discussion highlights some interesting challenges that new technology such as NVMe-oF is introducing into the data centre. As we move to a model of highly parallelised workloads, the interaction between storage and compute is back under the spotlight. Fungible is working on storage products today but claims to be able to disaggregate GPUs and in the future, potentially DRAM. Interesting times.

Find out more at https://www.fungible.com/

Elapsed Time: 00:49.25

Timeline

• 00:00:00 – Intros 00:01:30 – What problem is Fungible looking to solve? 00:02:45 – The benefits of Moore's Law growth are almost flat 00:03:35 – Modern applications are data-centric 00:05:00 – Fungible is working on disaggregated architectures 00:06:15 – There's storage for compute and storage for – storage! 00:07:45 – What's "hyper-disaggregation"? 00:09:30 – Fungible offers volume-specific characteristics like encryption 00:11:00 – Everything can be disaggregated except DRAM (for now) 00:13:10 – Data I/O has specific requirements including in-order processing 00:14:15 – Fungible can use an Ethernet network at 90% without packet drop 00:15:30 – Data-centric workloads are heavily multiplexed 00:18:45 – Does disaggregation finally deliver a real software-defined data centre? 00:20:00 – Mainframes "reconfigured" overnight for batch workloads 00:21:30 – Even hyperscalers operate in silos 00:24:15 – The Fungible Storage Cluster – SAN 2.0? 00:25:30 – New hero numbers! 15 million IOPS!! 00:30:45 – Volumes are virtual across any or all storage clusters 00:34:30 – Fungible claims better than local performance 00:35:00 – The Storage Cluster gains additional benefits with a host DPU 00:37:00 – Let's not get diverted towards VSAM! 00:39:20 – Disaggregated technologies could deliver truly reconfigurable data centres 00:41:00 – Fibre Channel networks divide their traffic across multiple SANs 00:42:00 – Who is using the Fungible technology? 00:45:00 – Fungible is working on I/O primitives for SQL databases* 00:49:00 – Wrap Up

Related Podcasts & Blogs

• 194 – ScaleFlux & Computational Storage Devices #190 – NVIDIA BlueField SmartNICs & DPUs #180 – SmartNICs – Pliops Storage Processor #177 – SmartNICs and Project Monterey #96 – Discussing SmartNICs and Storage with Rob Davis from Mellanox

Pradeep's Bio

Pradeep’s career includes founding Juniper Networks,

View Details

In this week's episode, Martin and Chris discuss Computational Storage with Tong Zhang, Chief Scientist and co-founder at ScaleFlux. Computational Storage devices add value to traditional NAND by offloading data processes directly onto the storage media. ScaleFlux offers two families of CSDs, including the CSD 2000 series, which implements inline data compression to improve endurance and logical device capacity.

In this conversation, Tong covers the benefits of using CSDs as well as some of the challenges of implementation. It's likely we will se CSDs being used for AI/Analytics pre-processing, especially in the public cloud.

More information on ScaleFlux can be found at https://www.scaleflux.com/ with further technical details of Computational Storage on the SNIA Website.

Elapsed Time: 00:43:39

Timeline

• 00:00:00 – Intros 00:04:00 – What is the ScaleFlux view of Computational Storage? 00:06:45 – What will the drivers of Computational Storage be? 00:08:40 – Compression can increase endurance and capacity 00:11:00 – The CSD2000 does inline compression/decompression in FPGA 00:13:05 – Why aren't all vendors doing inline compression? 00:14:45 – Databases make a good use case for CSD 00:17:10 – Compression can be 2:1 or as high as 5:1, depending on data 00:20:15 – SSDs do get hot! 00:21:30 – FPGAs will be replaced by ASICs in the future products 00:23:00 – AI/Analytics is a big target for Computational Storage 00:25:00 – Advanced functionality may require APIs 00:27:00 – NVMe will be used as a protocol to push code to CS drives 00:30:20 – Compute and storage are going to have to work closer together 00:35:00 – Computational Storage adoption will be evolutionary 00:38:00 – How will RAID/erasure coding be affected with CS? 00:40:20 – ScaleFlux will support PCIe-4/5 and possibly PLC NAND* 00:43:00 – Wrap Up

Related Podcasts & Blogs

• 190 – NVIDIA BlueField SmartNICs and DPUs #180 – SmartNICs – Pliops Storage Processor #177 – SmartNICs and Project Monterey* #96 – Discussing SmartNICs and Storage with Rob Davis from Mellanox

Copyright (c) 2016-2021 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #8vwo.

View Details

This week, Chris and Martin are talking to a podcast repeat offender, Subbiah Sundaram, VP of Products at HYCU. HYCU has recently announced the availability of Protégé for Microsoft Office 365, delivered as SaaS or Backup as a Service (BaaS). This continues an expansion of the HYCU and Protégé backup offerings that started with Nutanix data protection and has expanded past on-premises virtualisation to encompass the public cloud and now SaaS.

The conversation covers a wide range of topics relating SaaS and data protection, including the way in which services are implemented via APIs provided by SaaS vendors. With unlimited data storage, SaaS backup offers real opportunities for data mining and analysis.

To learn more about Protégé for Office 365, follow the link to https://www.hycu.com/tryhycu/. You can find more details on the support for Office 365 in the press release – here.

Elapsed Time: 00:47:15

Timeline

• 00:00:00 – Intros 00:03:00 – Vendors do not back up your SaaS service (for you) 00:04:45 – Office365 is a multitude of separate services 00:09:00 – Where does HYCU BaaS reside? 00:10:00 – The charging model for BaaS is different to on-premises offerings 00:11:25 – How does the customer monitor success/failure in SaaS offerings? 00:17:00 – Some businesses are happy with self-restore, others are not! 00:18:30 – What features do SaaS vendors offer to do backup? 00:20:00 – SaaS API users need to rethink how their services are built 00:22:00 – Centralised credentials management is key to delivering SaaS 00:25:50 – How can unlimited storage be justified and what does it mean? 00:29:00 – Can unlimited storage be abused? 00:31:50 – Can unlimited be used as deliberately limited (active deletion)?00:35:50 – Is BaaS a perfect tool for e-discovery? 00:42:55 – How will HYCU bring in the existing non-SaaS service together with SaaS? 00:44:20 – The Public Cloud is the right place to doing backup analytics* 00:45:30 – Wrap Up

Related Podcasts & Blogs

• 165 – Homogeneous Data Protection with HYCU #73 – HYCU – Data Protection for Hyper-Converged Infrastructure HYCU Announces GA of HYCU for Azure* Backup os Your Responsibility – Even in Public Cloud

Copyright (c) 2016-2021 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #tv7t.

View Details

This week, Chris and Martin chat to long-time friend and Kubernetes legend, Nigel Poulton. Nigel is well-known in the industry for producing training courses and books on Kubernetes, although was once a storage person in a previous life. The aim of this podcast episode is to examine how storage and Kubernetes come together. However, we start by asking Nigel to explain his pivot to containers and now the Kubernetes ecosystem.

This discussion touches on some interesting aspects of how persistent storage and Kubernetes should be managed together, which today is via the CSI (Container Storage Interface). Is this plugin a long-term solution for data mobility? We also manage to get an obligatory mainframe reference into the conversation.

Sadly we didn't get through all of our discussion topics in this long-running episode, so Nigel will be back later in the year to continue the conversation.

You can find more on Kubernetes, over at Nigel's Website: https://nigelpoulton.com/

Elapsed Time: 00:53:08

Timeline

• 00:00:00 – Intros 00:01:00 – Nigel is looking for things to "read" 00:05:10 – Why did Nigel pivot to Kubernetes? 00:07:50 – Is Kubernetes the future of containers? 00:09:55 – Kubernetes needs to avoid the OpenStack risk 00:13:15 – Are we building permanent or temporary cluster? 00:17:40 – Why did AWS open source EKS? 00:22:30 – Shouldn't we talk about storage now? 00:25:40 – Networking is just "pass the parcel" (hot potato) 00:28:00 – Customers should be using CSI-supported storage 00:32:00 – Nigel believes data mobility should be an application responsibility 00:35:10 – Obligatory mainframe reference (DFSMS) 00:39:20 – How should autoscaling work for storage and Kubernetes? 00:42:20 – Why is QoS in storage not seen more frequently? 00:49:00 – Nigel is into muscle cars* 00:51:00 – Wrap Up

Related Podcasts & Blogs

• 53 – Persistent Storage and Kubernetes with Evan Powell #151 – Introduction to StorageOS v2.0 #145 – Anthos Ready Storage for the Enterprise #129 – Choices for Persistent Container Storage with Niraj Tolia Will We Care About Kubernetes in 2025?

Nigel's Bio

Nigel's a technology geek, author of three utterly mind-blowing and life-changing books, and creator of weapons-grade Kubernetes training videos (his words not ours).

Copyright (c) 2016-2021 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #lx84.

View Details

This week, Chris and Martin are in discussion with Cathy Southwick, CIO at Pure Storage. The topic of conversation is the results of a CIO survey undertaken by Pure to look back at the challenges of the coronavirus pandemic and how customers have changed their IT strategies. The discussion covers how Pure Storage has adapted to the lack of in-person site visits and adapting to remote installation and operations. Digital transformation projects have continued, while customers have adapted their priorities and goals to align with the challenges presented by COVID-19. Public cloud has been an easy target for migrations. Customers are now looking at how best to rebalance workloads between on-premises and the public cloud as companies stabilise and re-adjust to a COVID-compliant way of working.

You can find Cathy's LinkedIn article and the results of the survey here – https://www.linkedin.com/pulse/what-we-can-learn-from-cios-covid-19s-impact-cathleen-southwick/

Elapsed Time: 00:32:51

Timeline

• 00:00:00 – Intros * 00:00:30 – Martin gets trolled by Google Maps! * 00:03:30 – How has Pure and their customers managed the Pandemic? * 00:06:00 – How has data centre access being managed? * 00:09:00 – Digital transformation projects have continued to be delivered * 00:14:00 – Automation is top of the priorities, with security and customer experience * 00:16:30 – Employee well-being figured highly * 00:17:30 – How have customers adopted the cloud? Tactical or Strategic? * 00:21:00 – FinOps – Financial operations for cloud will be the dream next job 00:24:00 – Will Cloud drive the "as a service" models? * 00:26:10 – The Portworx acquisition indicates a pivot towards data and data mobility * 00:28:35 – EMC – Where Information Lives – describes the future for storage companies * 00:30:40 – Wrap Up

Related Podcasts & Blogs

• 152 – Post Pandemic Storage Efficiencies #149 – Coronavirus 2.0 #146 – Coronavirus and Impacts on the Technology Industry* #185 – Pure-as-a-Service 2.0

Cathy's Bio

Cathy Southwick joined Pure Storage in 2018 as Chief Information Officer. In this role, she leads Pure’s global IT strategy and advances the company’s operations through the delivery of next-generation technology capabilities and systems.

Cathy is an accomplished leader with over 20 years of experience defining and executing forward-looking IT strategies. Prior to Pure, Cathy held leadership positions at AT&T, including Vice President, Technology Engineering and Vice President, Cloud Planning & Engineering. During her tenure at AT&T, Cathy led the planning and execution of IT strategies from the Core Network, IT application modernisation, and the IT cloud.

Before joining AT&T, Cathy spent 11 years at Viking Freight System (now owned by FedEx) where she held escalating leadership positions in IT architecture and planning, software development, merger integration, strategic planning, human resources management, procurement, project/portfolio management, and process re-engineering.

View Details

This week, Chris and Martin chat to Kevin Deierling, SVP of Marketing for Networking Products at NVIDIA. SmartNICs and DPUs (Data Processing Units) are starting to become mainstream as application use-cases such as AI and analytics drive a need for greater data throughput and performance. Kevin explains the design and thinking behind BlueField, NVIDIA's family of DPU products that combine offloaded network, storage and security functionality.

Elapsed Time: 00:47:36

Timeline

• 00:00:00 – Intros 00:01:30 – We're not a networking podcast! 00:02:15 – Why will we need DPUs and SmartNICs? 00:04:30 – Von Neumann is diverging 00:06:50 – What is the BlueField architecture? 00:08:45 – A DPU could act as a storage array controller 00:11:30 – Storage DPUs make devices appear local 00:13:00 – DPUs enable efficient bare-metal server deployments 00:15:00 – Storage, networking & security use around 30% of traditional cores 00:16:00 – Does a DPU represent better or worse performance than CPU? 00:20:15 – NVIDIA DPUs emulate existing devices, reducing application changes 00:25:00 – Where does the outboard management take place? 00:26:00 – DOCA is the application framework for DPUs 00:28:00 – BlueField 2X combines GPU and DPU on the same card 00:31:40 – DPUs enable the real-time nature of data processing 00:35:40 – Mainframe reference! 00:38:00 – Where will initial adoption take place? 00:40:00 – How does the use of SmartNICs affect TCO? 00:45:10 – The future is 1000x improvement with BlueField 4* 00:46:00 – Wrap Up

Related Podcasts & Blogs

• 96 – Discussing SmartNICs and Storage with Rob Davis from Mellanox #177 – SmartNICs and Project Monterey #180 – SmartNICs – Pliops Storage Processor VMware Project Monterey – First Impressions Fixing the x86 Problem

Copyright (c) 2016-2021 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #ky1r.