Storage Unpacked Podcast

This week, Chris and Martin discuss the availability of super-high capacity or ultra-capacity flash drives.  The conversation comes out of an announcement from Nimbus Data that has produced a 100TB SATA SSD.  This follows up from last year’s 50TB drives that were OEMed to Viking and Smart Modular.  You can find some more background at Architecting IT here.  With so much capacity in a single 3.5″ form factor that could cost upwards of $50,000, is this product practical?  How do the drives survive failure?  Can they be repaired and who would build systems from them?  One of the most important questions to answer is the risk factor of putting so much content into a single device and the impact of theft to the end user.  All fun topics!  As Chris and Martin say, if you want your 30+TB drive reviewed, just send a few over… Elapsed Time: 00:22:31 Timeline * 00:00:00 – Intro * 00:01:00 – 100TB SATA SSD from Nimbus * 00:01:30 – What are the practicalities – failure domain? * 00:03:30 – Can SSDs be repaired? * 00:05:50 – How many enterprises have 100TB of active data? * 00:06:35 – What size array could you build, minimum drive count? * 00:08:08 – RAID rebuilt and throughput times, does this represent a risk? * 00:11:00 – Do the economics of keeping spare $50,000 drives make sense? * 00:12:30 – What is the target market?  Hyper-scalers? IoT? * 00:13:30 – Remember memory chips being stolen from servers? * 00:15:00 – Could this technology be used for smaller, compact drives? * 00:15:30 – What about Ruler and NGSFF devices? * 00:18:45 – When will we have 200TB drives? * 00:20:30 – Another use case – content delivery. * 00:21:20 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

This week, Chris and Martin review the all-flash market with Chris Mellor.  We looked at this sector around 12 months ago, so we’re ready for an update.  Lots of things have happened over the past 12 months.  Vendors have gone away and come back (Violin).  Existing vendors have seen a resurgence in their sales (IBM and NetApp), while others have struggled.  Pure Storage is now an incumbent alongside the likes of Dell EMC, IBM, NetApp and HPE. The conversation moves on to discuss some interesting start-ups like WekaIO and both Excelero and E8 Storage (previous guests).  NVMeoF comes back as a recurring theme.  The podcast wraps up with a discussion of the Chinese influence and whether things are afoot with Western Digital and Cisco.  Could 2018 be the year Cisco acquires NetApp? Elapsed Time: 00:32:57 Timeline * 00:00:00 – Intro * 00:01:15 – A quick review of some changes in the market * 00:01:20 – Violin goes and comes back, IBM, NetApp emboldened. * 00:02:00 – IDC 4Q2017 all-flash figures, any surprises? * 00:03:50 – Where has Hitachi gone? * 00:04:40 – NVMe & NVMeoF – everybody’s doing it. * 00:06:00 – Are there really any options for new “traditional” flash arrays? * 00:08:00 – Pure Storage – what was their market strategy? * 00:09:00 – Back to NVMeoF! * 00:11:20 – What about all-flash distributed file systems? * 00:11:40 – Whoops – you mean Exelero, not E8, Chris! * 00:13:40 – WekaIO and SPECSFS2014 results * 00:16:38 – Toshiba – who knew? (Chris apparently) * 00:18:00 – are M.2 drives really “enterprise”? * 00:20:00 – What about the Chinese companies like Huawei? * 00:22:30 – Is there really a dedicated all-flash market any more? * 00:24:00 – Could we see an HCI 2.0 emerge, like the Datrium architecture? * 00:25:00 – How much demand will there be from ML/AI? * 00:28:30 – Is something going to happen with Western Digital & Tegile? * 00:29:00 – Cisco to acquire Pure? (or NetApp) * 00:32:00 – Wrap up! Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

In this week’s episode, Martin and Chris talk to Fred Monsone, founder of A3 Communications, a boutique storage PR and communications company.  It’s clear that modern day PR isn’t the right term any more and communications describes the work much more accurately.  Fred describes the reasons why start-ups and seasoned companies want to find the right way to engage with the press & journalists, bloggers and other influencers.  Dealing with people is a challenge, especially the divas (like Martin), who expect bacon rolls at events!  In case you’re wondering, we mentioned A3 Technology Live, an event that Fred runs and Martin & Chris have attended in the past.  Finally the conversation wraps up with a few horror stories and a problem with steamy windows! Elapsed Time: 00:38:23 Timeline * 00:00:00 – Intro * 00:01:00 – Is it PR or Communications – what’s the difference? * 00:02:20 – Who do comms companies talk to? * 00:04:00 – First hand experience counts in being an influencer * 00:05:00 – Who are the clients – startups? Incumbents? * 08:50:00 – Why choose storage as a specialist PR company? * 00:11:00 – How has the storage market grown over the last 15 years? * 00:13:00 – Are there any EMEA storage bloggers left? * 00:15:30 – I’m selling a new storage product – how does PR work? * 00:18:45 – How do the American and European markets differ? * 00:22:00 – Are you a “leading” provider? * 00:23:00 – How can you measure coverage and success of PR? * 00:28:00 – Being selective about reading interesting influencers. * 00:31:30 – Quirky people, dealing with the divas! (including bacon rolls) * 00:33:00 – Vendor horror stories, spill the beans! * 00:35:00 – Steamy windows! * 00:37:00 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

This week, Martin, Chris and Gavin reflect on the decision by EMC to disband the {code} team and discuss whether open source has a place in storage.  {code} was a project started by EMC before the Dell acquisition. The team focused on open source advocacy and developing tools such as storage plugins for Docker.  Looking wider, companies like Red Hat have built their business on open source, but how are the storage platforms working out?  Can vendors make money from open source and do we need a benevolent dictator like we have in Linux? Elapsed Time: 00:23:18 Timeline * 00:00:00 – Intro * 00:00:30 – The demise of EMC {code} * 00:03:00 – Does open source have a place in storage? * 00:06:00 – Open source storage hasn’t come from existing companies * 00:07:30 – Is open source (storage) simply a loss leader for other business? * 00:08:00 – What about SAMBA?  The biggest open source storage platform? * 00:08:30 – ViPR, iWave and open sourcing the control plane. * 00:11:00 – Is open source storage good for storing data we can afford to lose? * 00:13:30 – Paid storage lets customers play the blame game * 00:15:35 – So how can vendors make money from open source storage? * 00:18:00 – ScaleIO – interesting usage model (although this has changed) * 00:20:00 – Did storage need a benevolent dictator like Linus? * 00:21:33 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

This week’s podcast was recorded a few months ago, so apologies for the sound quality.  The team discuss the role of the storage administrator and whether we will see the eventual demise of storage admins or simply a transition to another role.  The conversation tracks the responsibilities of storage administrators over the last 30 years, from the era of the mainframe to today’s software-defined offerings. Early on, Martin invents a new word (abomnerbation)!  The discussions finally agree that the age of the LUN monkey is over. Elapsed Time: 00:26:57 Timeline * 00:00:00 – Intro * 00:01:00 – Storage on the mainframe – including punched cards, 300GB under management. * 00:03:00 – Storage Admin in the 1990s – VSS and ESS * 00:04:00 – Centralised storage (EMC) introduced to fix the mess of Windows * 00:04:33 – StorageTek Iceberg, or RAMAC Virtual Array * 00:05:20 – EMC horror stories, Gavin – spill the beans! * 00:07:00 – Storage was very “hands on” – Bus & Tag! * 00:08:30 – 2000’s introduced the move towards simplicity and awful SRM products * 00:10:00 – Bring on the APIs! * 00:12:00 – Reliability – one of today’s big improvements * 00:13:00 – How reliable were those old mainframe disks? * 00:14:00 – Who made things easier – XIV and 3PAR. * 00:15:30 – Storage is now highly automated for performance. * 00:16:30 – Hardware costs have declined, why pay for an expensive human resource? * 00:17:30 – Using policy to determine how storage should be managed * 00:19:00 – The storage role is changing to architectural and design – with less understanding? * 00:22:00 – Who will do low level support in the future, the vendor? * 00:24:00 – Start to understand the value of the data you are managing. * 00:26:00 – The LUN Monkey is dead!! * 00:26:08 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

In this week’s podcast, the guys talk about some of the myths in the storage industry.  The focus is mainly on traditional storage arrays and some of the features that end users might not need.  The conversation takes a marketing turn, with questions on how five 9’s availability is marketed and the use of dedupe/compression as a tool to reduce costs. What do you think? Are there any myths you believe exist within the industry?  Drop us a line and give us your thoughts or areas we should be discussing. Elapsed Time: 00:31:05 Timeline * 00:00:00 – Intro * 00:01:00 – Myth 1: Do I really need synchronous replication? * 00:07:13 – Myth 2: Snapshots are good enough for backup. * 00:09:40 – Myth 3: Can I really have a five 9’s storage array? * 00:14:10 – Myth 4: Non-disruptive upgrades? * 00:19:42 – Myth 5: Are performance figures really accurate? * 00:27:20 – Myth 6: Effective capacity claims by vendors. * 00:30:00 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

In this week’s podcast, the team talk to Don Foster, Senior Director for WW Solutions Marketing at Commvault.  Initially this podcast was intended to be recorded at Commvault Go last November (2017) in Washington DC, but schedules didn’t align so we did it remotely.  The audio quality is not as good as we’d like, so apologies for that.  However, there are some good discussions in this episode, including lots to think about if you’re looking to deploy a backup appliance in place of a traditional DIY solution.  Don spends some time at the end of the podcast talking through the Commvault offering in this space – HyperScale.  Finally we wrap up with thought about how this market will mature.  Is there any future at all for build-your-own solutions? Elapsed Time: 00:33:13 Timeline * 00:00:00 – Intro * 00:01:00 – Commvault background * 00:04:00 – Why buy an appliance compared to build your own? * 00:06:30 – What operationally, is the benefit for the customer? * 00:07:30 – Vendor designs, scale-out or scale-up? * 00:08:30 – What are the compromises? * 00:10:00 – Who are the typical customers for the appliance model? * 00:11:30 – Are appliances simply a way to make better margins? * 00:14:30 – Design and build is over…. * 00:15:30 – Nobody loves their 10th NetApp – allegedly. * 00:17:45 – Deeper dive into Commvault HyperScale. * 00:22:00 – What patching has been done for Spectre/Meltdown? * 00:24:30 – How are existing non-appliance customers accommodated? * 00:26:01 – What support is available for public cloud? * 00:29:30 – How is licensing managed? * 00:30:30 – What is the future of data protection appliances? * 00:32:30 – Wrap up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

This week Chris and Gavin catch up with Chris Mellor, Storage Editor for The Register.  With so much happening in storage, it’s difficult to know where to start, so the guys focus on the rumour of Dell EMC reversing into VMware.  Could this really happen and why is Dell EMC even thinking about this?  The conversation flows on to IBM and their super-mega-hyper-uber announcement on NVMe.  Is IBM getting its storage mojo back?  Finally, the discussion turns to IoT and how storage and compute move to “The Edge” – no, not Dave Evans, but IoT.  What can we expect from new companies like ScaleFlux and NGDS?  The wrap up puts Chris on the spot for his predictions and gives us a nice follow-up in 6 months time. Elapsed Time: 00:27:02 Timeline * 00:00:00 – Intro. * 00:00:37 – Straight into Dell and the VMware reverse IPO. * 00:02:31 – IDC numbers for 3Q2017 didn’t favour the big boys. * 00:03:40 – Is this actually the start of a change in Dell EMC’s business? Solution Stacks! * 00:04:50 – EMC chopped, shaped and distributed. * 00:06:40 – IBM Storage reinvigorated with NVMeoF? * 00:08:00 – Is 2018 the year for NVMe over Fabrics? (Or VDI, or Linux Desktop) * 00:09:40 – But how big is the NVMeoF market, is it still niche? * 00:12:00 – How does high the performance technology translate to Edge/IoT? * 00:13:30 – What is in-situ processing? ScaleFlux, NGDS (Next Generation Data Systems) * 00:15:30 – How important will compute next to storage be? * 00:17:00 – Is Edge/IoT too hard, when simple issues like security never gets fixed? * 00:19:00 – What would the more practical uses of IoT be? * 00:21:00 – Could IoT be like Cryptocurrency or Bitcoin? * 00:22:30 – What is the future of Storage? Latency, NVMeoF and SCM… * 00:26:20 – Wrap Up Vendors Mentioned in this Podcast Dell EMC, VMware, IDC, NetApp, IBM, E8 Storage, Lenovo, HPE, Toshiba, Pure Storage, Kaminario, Violin Systems, DSSD, Excelero, FusionIO, X-IO

In this week’s episode, the team talk to Rob Peglar, Senior VP and CTO at SymbolicIO.  The conversation covers persistent memory and in particular the NVDIMM format of devices that provide persistent storage in a standard server DIMM slot.  Rob explains how the standards are set through JEDEC, with a description of the three types – NVDIMM-N, F and P. The discussion continues to use cases, which seem mainly to be for any application needing low latency with local server persistence.  Databases and analytics are the main beneficiary, however Rob believes that the hyper-scalers may be able to take advantage of the technology in the future. A few references are made during the podcast.  The SNIA Persistent Memory sub-site can be found here at https://www.snia.org/pm.  Here’s a link to the Persistent Memory Summit – https://www.snia.org/pm-summit.  It’s actually Linux kernel 4.2 onwards that adds NVDIMM support as well as Windows 2016.  This Microsoft webinar provides some detail on how NVDIMM can be used with Storage Spaces Direct – Using Non-volatile Memory (NVDIMM-N) as Block Storage in Windows Server 2016. Elapsed Time: 00:28:46 Timeline * 00:00:00 – Intro * 00:00:54 – Persistent Memory defined * 00:02:00 – Why use it? What is it for? * 00:04:40 – Why not just use a honking great battery for protection? * 00:05:20 – Standards from JEDEC * 00:06:00 – Defining the NV-DIMM formats. N, F and P. * 00:12:00 – Could 3D-XPoint justify its own definition? * 00:13:35 – How are these devices supported by the hardware? * 00:16:00 – What are the application use cases for persistent memory? * 00:19:00 – Could I trust having NVDIMM as my only source of data? * 00:23:00 – Who is buying and using NVDMM today? * 00:27:30 – Wrap up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.  

In this week’s podcast, the team talk to Alex Chircop, CTO at StorageOS about the implications on storage of the recent Spectre/Meltdown vulnerabilities.  Much has been made of the potential impact to I/O performance and by definition storage platforms and products.  The guys talk about what the vulnerabilities actually mean for end users and what to expect from storage vendors.  Finally, the podcast concludes with some suggestions from Martin, as the token end user in the discussion. Two references were made in the podcast.  The first is to a Techspot article comparing NVMe and SSD performance.  You can find that here.  The second was the reference to an inpromptu Slack group, set up to work on patches.  You can find that here. Elapsed Time: 00:25:17 Timeline * 00:00:00 – Intro * 00:01:00 – Spectre/Meltdown vulnerabilities defined * 00:05:30 – How are patches affecting the workload of applications? * 00:09:30 – How successful have existing patches been? * 00:10:30 – Was the awareness of the issue handled correctly? * 00:12:00 – Has the industry overreacted to the problems? * 00:13:00 – How are storage products affected? * 00:17:30 – What should end users and storage vendor customers do? * 00:21:00 – NVMe and all-flash may suffer more than spinning media systems. * 00:24:00 – Wrap Up Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

In this week’s podcast, the team chat again with Jim Handy from Objective Analysis.  Jim is an analyst covering the semiconductor industry and provides some great insight into where pricing is headed in the NAND flash market.  The conversation covers some of the technical challenges the industry faces, as well as the impact from China entering the market.  Finally, the guys discuss the impact on Enterprise storage pricing. There are some interesting observations in this conversation, such as the capability for NAND to be used in multiple densities and how the ability for the Enterprise market to follow the consumer market is driven by increasingly more sophisticated controllers.  Enjoy! You can find out more about Jim’s company at http://objective-analysis.com/ and follow him on his two blogs at http://thessdguy.com/ and http://thememoryguy.com/. Elapsed Time: 00:29:26 Timeline * 00:00:00 – Intro. * 00:01:45 – Why has there been delays in getting 3D-NAND to market? * 00:03:00 – MLC, QLC, TLC, XTC, do the users care? * 00:04:54 – Stupid question time, can NAND morph into any type of media? * 00:06:30 – Consumer leads the enterprise towards cheaper NAND products. * 00:07:30 – Who gets their NAND first? * 00:09:20 – Let’s talk about China and $140 billion! * 00:15:50 – Where can we expect the price of NAND and DRAM to go in the long term? * 00:19:00 – How far off are we from the next price collapse? * 00:21:30 – How will price drops affect the Enterprise market? * 00:28:30 – Wrap Up! Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission.

In this podcast, recorded in December 2017, Chris talks to Nigel Tozer, Solutions Marketing Director, EMEA at Commvault.  Today the subject is GDPR, otherwise known as the General Data Protection Regulation.  GDPR is an update to existing data protection legislation that is meant to reflect the needs of EU citizens in protecting their personal data in a digital age. Chris and Nigel talk through the issues, some of the steps end users can take and how Commvault offers solutions to bring businesses in line with GDPR. Elapsed Time: 00:27:52 Timeline * 00:00:00 – Intro * 00:01:40 – What is GDPR? * 00:03:40 – Why is GDPR such a big deal now? * 00:05:00 – Will GDPR have the bite to match the bark? * 00:07:44 – How have the rules been tightened in the UK compared to previous data protection? * 00:09:00 – Reduce the legalise, please! * 00:10:00 – Portability and ownership of data, how will it work? * 00:12:00 – It’s people, process and technology. * 00:13:00 – What practical changes need to be made? * 00:15:00 – Right to be forgotten versus the law – who wins? * 00:17:19 – What about technology? * 00:20:30 – The positive side of GDPR * 00:23:20 – Will people change their behaviour? * 00:27:00 – Wrap Up In the podcast, Nigel mentioned the ICO statement regarding the Uber breach.  More details are here (The Independent).  For more details on Commvault’s GDPR solutions and advice, check out www.commvault.com/gdpr.

In this podcast, recorded on 21 November 2017, Chris, Gavin and Martin talk about thoughts from Chris’ recent trips to CommVault GO, Storage Field Day 14 and NetApp Insight.  The CommVault and SFD14 events were held in the US, with NetApp Insight talking place in Berlin, Germany.  At Storage Field Day, Dell EMC, E8 Storage, Scality and Congruity 360 presented.  This podcast talks about the Dell EMC presentations. E8 Storage has been a recent podcast guest. Feelings on the events and content were mixed.  Some good and some bad.  What do you think?  Leave us your feedback as a comment or on Twitter.  You could even use our LinkedIn group. Elapsed Time: 00:30:48 Timeline * 00:00:00 – Intro * 00:01:00 – CommVault – great data protection, what else do they do? * 00:03:00 – GDPR – the world is going to end!! * 00:05:00 – Robert Swan – 2041.com, trek to the South pole * 00:06:00 – Apologies in advance to all marketing people… * 00:06:30 – Storage Field Day 14 and Dell EMC. * 00:07:00 – The benefit of SRDF on VMAX and XtremIO X2. * 00:08:00 – Back to the data services question again * 00:11:00 – Dell EMC Midrange discussion * 00:13:30 – Dell EMC Backup solutions. * 00:14:30 – The marketing hype of Ransomware * 00:16:00 – Creating consistency across backup solutions – please! * 00:17:00 – Everybody loves an EBC * 00:18:30 – NetApp Insight, moving on from Las Vegas to Berlin * 00:20:00 – Differentiating HCI offerings * 00:25:00 – Railing on Microsoft * 00:27:00 – Looking ahead to AWS re:Invent * 00:28:30 – Gartner infrastructure conferences * 00:29:00 – Bananarama and Rick Parfitt Jr Band * 00:30:30 – Wrap up Companies & Products Referenced Equifax, CommVault, Dell EMC, VMAX, XtremIO, Equallogic, Dell SC, Unity, VNX, HPE, NetApp HCI, SolidFire, Nutanix, AHV, Microsoft Azure, NFS for Azure. Copyright (c) 2016-2017 Storage Unpacked.  No reproduction or re-use without permission.  

In this podcast, recorded at NetApp Insight in Berlin, November 2017, Chris talks to Martin Cooper, Senior Director of Field Engineering, Next Generation Data Centre.  The conversation introduces HCI, NetApp’s hyper-converged offering, which recently went GA.  NetApp HCI uses SolidFire storage and scales differently to traditional HCI solutions in that the processor cores aren’t shared by storage and compute.  Instead each is delivered as a separate server, more like CI. Martin and Chris discuss the reasons for this architectural choice and how it fits a different market space than traditional hyper-converged solutions. Elapsed Time: 00:11:54 Timeline * 00:00:00 – Intro * 00:00:40 – Crowd control at the HCI Booth! * 00:01:00 – HCI – first internally developed product for some time (FlashRay?) * 00:02:45 – FlexpodSF – Flexed on Cisco using ElementOS * 00:03:30 – NetApp HCI, how is it differentiated? * 00:06:00 – Exploiting the SolidFire architecture – programmable storage * 00:06:30 – Day one support – vSphere – but it could be used for other things * 00:09:00 – Integrating into the “Data Fabric” story – the infrastructure fabric * 00:11:00 – Wrap Up Links * NetApp HCI homepage * NetApp Blogs (filter middle option on HCI) Copyright (c) 2016-2017 Storage Unpacked.  No reproduction or re-use without permission.

This podcast was recorded at NetApp Insight in Berlin on 15 November 2017.  Chris talks to Dave Hitz, founder of NetApp Inc.  The conversation is fairly free flowing, covering a general discussion on data management, NetApp’s positioning with Azure NFS and how we can expect to see hybrid cloud work in the future.  You’ll notice there’s no discussion on any hardware products at all and little product discussion of any kind. This conversation could have continued for hours (and in fact we did continue off-air), but we kept it to 20 minutes. Elapsed Time: 00:19:05 Timeline * 00:00:00 – Intro * 00:00:30 – Embracing the cloud * 00:01:30 – Azure Files, NFS?  What’s it called? * 00:03:30 – Linux in Azure – 30% or 40% * 00:05:00 – Transforming into data management * 00:06:00 – The stickiness of cloud – additional functionality creates a pull * 00:08:20 – What is big data?  Big files, big downloads or lots of complex content? * 00:09:30 – Yes, the UK is still part of the EU for now and we will have GDPR * 00:11:00 – Data security, how do we fix it – is the cloud more or less secure? * 00:12:00 – Public reporting of data breaches, Sunshine laws * 00:13:00 – How long will we have a hybrid world? * 00:15:00 – Three ways to build IT * 00:17:30 – Will NetApp become a service provider? * 00:19:00 – Wrap up Copyright (c) 2016-2017 Storage Unpacked.  No reproduction or re-use without permission.