Technology:Unpacked

This week we talk to Mark Carlson, co-chair of the SNIA Technical Council, about the storage needs of hyperscalers.  Mark defines hyperscalers as those companies opening multiple data centres a year, most notably Amazon, Microsoft Azure, Google and Facebook in the US and Baidu, Alibaba and Tencent in China.  These vendors are deploying petabytes of […]

This week, Chris and Martin discuss the subject of price parity between all-flash and hybrid or HDD-based storage arrays.  From Martin’s perspective, he is starting to see vendor pricing getting close to parity and it making more sense to buy all-flash than a spinning media device.  However, what are the issues?  TCO is one – it’s not all about array pricing.  Also, flash offers more opportunity to be Opex focused, as it’s easier to increment flash in an array than adding many disks to maintain performance.

So what about hybrid?  Does it have a position, if the marketing is going all-flash for traditional workloads and all-HDD for archive and unstructured data?  The conversation closes on questioning whether vendors offering rental rather than buy makes sense and whether the industry can afford not to go all-flash.

During the podcast, we mention Chris Mellor’s article on flash revenue.  You can find it here.  We also talk about the Micron 5120 QLC SSD.  You can find some thoughts on that product here.  Finally, we talk about the 40-year anniversary Intel x86 processor.  You can find more details here.

Elapsed Time:  00:25:58

Timeline

* 00:00:00 – Intros

* 00:01:00 – Are we at price parity between all-flash and disk-based systems?

* 00:01:30 – What’s being pushed out?  15K, 10K drives?

* 00:02:30 – On what basis are we seeing price parity, raw, usable/effective?

* 00:03:30 – TCO is important, flash provides better environmental costs.

* 00:05:00 – Licensing – should you have to pay for licences you will never use? (discussed on #39)

* 00:06:30 – How is the price for flash declining? (discussed on #9)

* 00:08:20 – The 100TB flash drive. (discussed on #44)

* 00:09:00 – With flash, do we need tiering?

* 00:10:00 – Will we have a spectrum of flash or will SLC and MLC disappear?

* 00:14:00 – Could flash drives actually last much longer than the vendors claim?

* 00:14:50 – Backblaze data, making choices from available information.

* 00:18:00 – Are we at a time when a flash-first policy is realistic?

* 00:19:00 – Is there any point to hybrid arrays?

* 00:20:30 – Now that prices are in parity, would an Opex solution be more appropriate?

* 00:22:00 – Does flash make it easier to buy on demand and expand at smaller increments?

* 00:24:00 – Can we afford not to move to a flash solution with todays performance needs?

* 00:25:00 – It’s a Wrap

Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 934E.

This week, Chris and Martin discuss the subject of price parity between all-flash and hybrid or HDD-based storage arrays.  From Martin’s perspective, he is starting to see vendor pricing getting close to parity and it making more sense to buy all-flash than a spinning media device.  However, what are the issues?  TCO is one – […]

This week Chris and Martin talk to Evan Powell, CEO of OpenEBS and formerly the founding CEO of Nexenta.  The conversation covers the use of persistent storage with container orchestration tool, Kubernetes.  Despite what the industry might think, persistent storage that can be mapped to a container (or in this case pod) is still an important problem to solve.

Evan sets the scene for us with some background on pods, stateful sets, claims and storage classes.  As the conversation proceeds, the team discusses the way in which developers expect to consume cloud-native storage and in particular the abstraction of infrastructure.  Evan provides some extra background on CSI (Container Storage Interface) and how it is allowing storage vendors to interface with Kubernetes without lots of code changes.  Finally the wrap up covers what gaps are missing in current development and the projects Evan is working on.

You can find more information on OpenEBS at https://www.openebs.io/, with links to MayaOnline and Litmus, which Evan references at the end of the recording.  The discussion also mentions the CNCF or Cloud Native Computing Foundation, under which Kubernetes has been developed.  Evan makes reference to the Intel Software Performance Development Kit and to Google Fuchsia.

Why not give us some feedback?  You can find us on LinkedIn here.

Elapsed Time: 00:27:51

Timeline

* 00:00:00 – Intro.

* 00:01:00 – How is persistent storage connected to Kubernetes?

* 00:03:00 – Pods, stateful sets, claims, what does it all mean?

* 00:05:30 – Defining storage as code with storage classes.

* 00:06:30 – Static versus dynamic provisioning.

* 00:08:00 – Application or infrastructure-level data resilience?

* 09:00:00 – Developers, developers, developers!

* 00:11:00 – CSI – Container Storage Interface.

* 00:14:30 – Does CSI have enough industry knowledge?

* 00:16:00 – The storage industry has a terrible management software reputation.

* 00:17:00 – Is there an opportunity to bypass traditional storage protocols?

* 00:22:00 – Hurrah – Mainframe!  System Managed Storage.

* 00:23:00 – So where are the gaps, what’s missing?

* 00:26:00 – What is OpenEBS. Maya online and Litmus?

* 00:27:00 – Wrap Up.

Evan’s Bio

Evan Powell is CEO of MayaData where he helped found the popular OpenEBS containerized storage project.  Previously he has been founding CEO of three enterprise infrastructure software companies, Clarus Systems (RVBD), Nexenta Systems, and most recently StackStorm (BRCD).

Under his leadership Nexenta became the leader of the OpenStorage movement and the software defined storage sector, with thousands of customers and hundreds of millions of dollars of annual partner sales.

As founding CEO of StackStorm, Evan helped to create the event driven automation category, supporting a vibrant open source community leveraging and improving upon approaches used by the largest operators such as Facebook.   Prominent StackStorm users include Netflix, Cisco, and Dimension Data.  Brocade purchased StackStorm in 2016.

Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode JJ39.

This week Chris and Martin talk to Evan Powell, CEO of OpenEBS and formerly the founding CEO of Nexenta.  The conversation covers the use of persistent storage with container orchestration tool, Kubernetes.  Despite what the industry might think, persistent storage that can be mapped to a container (or in this case pod) is still an […]

This week’s guest episode was recorded live in Silicon Valley at the offices of WekaIO.  The company has developed a scale-out parallel file system called Matrix that was specifically designed to exploit NVMe storage and new fast networking.  Chris is onsite to talk with CEO and co-founder, Liran Zivbel.  Martin is dialled in remotely from the bowels of the Storage Unpacked offices.

The conversation covers how Matrix was developed to work with new media and at the same time address some of the issues seen in the use of parallel file systems such as managing high levels of throughput and small file content.  Building a scale-out file system is challenging, and Weka developed their own user space storage operating system to ensure the performance of NVMe could be fully exploited.  Liran discusses the SPEC SFS 2014 benchmark tests, which can be found here.  Matrix is available for AWS on i3 EC2 instances.  More details can be found here (PDF).

Elapsed Time: 00:27:27

Timeline

* 00:00:00 – Intros

* 00:01:00 – What is WekaIO Matrix?

* 00:02:20 – Why are parallel file systems hard to build?

* 00:04:30 – What is it that end users want from a parallel file system?

* 00:05:00 – How is Matrix packaged and delivered?

* 00:07:00 – A real-time userspace storage operating system.

* 00:08:00 – Native Linux and a native file system driver.

* 00:11:00 – Snap to object – store snapshots on an object store.

* 00:13:00 – How is data protected within Matrix?

* 00:16:00 – How does performance scale with additional hardware?

* 00:17:00 – How does Matrix deploy in public cloud?

* 00:18:00 – Matrix SPEC SFS benchmark results.

* 00:20:30 – What are the typical customer deployment models?

* 00:22:00 – Storage pets versus storage cattle.

* 00:24:00 – Consumption models.

 Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode F7G4.

This week’s guest episode was recorded live in Silicon Valley at the offices of WekaIO.  The company has developed a scale-out parallel file system called Matrix that was specifically designed to exploit NVMe storage and new fast networking.  Chris is onsite to talk with CEO and co-founder, Liran Zivbel.  Martin is dialled in remotely from […]

This episode was recorded at Pure Accelerate in San Francisco on Tuesday 22nd May 2018 just before the event started.  Chris talks to Matt Leib and eventually Ray Lucchesi, initially about the role of the reseller, then eventually moving towards NVMe over Fabrics.  As with all live recorded podcasts, there’s a little background noise, including the weekly noon siren test.  The group speculates on what we can expect to be announced by Pure, the benefits of the Evergreen program, then whether we’re headed back to a hardware future for storage.

In the recording Ray talks about a recent vendor interview.  We’re fairly certain he’s talking about Apeiron.  Look out for a follow-up to this podcast, discussing what was actually announced by Pure Storage.

Elapsed Time: 00:19:25

Timeline

* 00:00:00 – Intros.

* 00:01:30 – The role of the reseller at tech conferences.

* 00:04:00 – What is the rate of technology adoption really like?

* 00:07:00 – Argh!  Single Pain of Glass!!

* 00:08:00 – What about NVMe adoption?

* 00:08:45 – Will Pure release new technology?

* 00:10:00 – Evergreen – good for Matt’s customers.

* 00:11:30 – Ray joins us for a discussion of NVMe over Fabrics (and Apeiron)

* 00:13:30 – And we get an air raid warning test!

* 00:15:00 – What comes after NVMeoF?

* 00:16:00 – Are we headed back towards a hardware future for storage again?

* 00:17:00 – What is the impact for Software Defined Storage?

Copyright (c) 2016-2018 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 2G0J.

Disclaimer: Chris was an invited guest at Pure Accelerate.  Pure Storage covered flights and accommodation costs.  There is no incentive or requirement to produce media from the event and no review of any content is made by Pure Storage before publication.  Pure Storage Inc is a client of Brookend Ltd.

This episode was recorded at Pure Accelerate in San Francisco on Tuesday 22nd May 2018 just before the event started.  Chris talks to Matt Leib and eventually Ray Lucchesi, initially about the role of the reseller, then eventually moving towards NVMe over Fabrics.  As with all live recorded podcasts, there’s a little background noise, including […]

How does quantum computing relate to storage?  This week’s episode is a bit of a diversion from our normal podcasting topics.  In episode #47 (Enterprise Storage is not Boring), we mused as to how quantum computers use data.  With no real knowledge between us, we asked for listeners to help us out.  As a result, we’re joined in this episode by Scott Crowder, Vice President & CTO, Quantum Computing, Technical Strategy & Transformation, IBM Systems.  That’s a long job title, Scott!

During the show we try and get to the bottom of what quantum computing really means.  What niche does it serve?  What does a quantum computer look like?  How do I program it?  We also ask the ever present question on encryption and whether quantum will break all of our secure transmission protocols.

As for the question on data, it seems quantum computing is more mundane in the storage area than expected.  We can expect classic computers to be around for a long time to do the traditional number crunching.

The podcast references number of items, including: vector co-processors for mainframe, Grover’s Algorithm, rotating wave approximation, Shor’s Algorithm and of course links to IBM’s Q Experience and the Q Network.  There is also some additional background on quantum computing from IBM here and you can try out the card test game here, which shows how quantum processing differs from classical computing.

Elapsed Time: 00:27:01

Timeline

* 00:00:00 – Intros.

* 00:00:30 – What is Quantum Computing?

* 00:02:40 – Quantum Mechanics – Superposition and Entanglement.

* 00:05:00 – It’s all about solving exponentially hard problems.

* 00:05:16 – Some examples, factoring, quantum chemistry, travelling salesman.

* 00:07:00 – What does entanglement mean?

* 00:08:00 – What is a qubit?

* 00:11:25 – What does a quantum computer actually look like?

* 00:13:00 – Programming with Python! With a classic and quantum mix.

* 00:15:00 – But how do we actually measure the result of an execution of code?

* 00:19:00 – How does this relate to storage, data and practical applications?

* 00:22:00 – How will Quantum Computing impact encryption?

* 00:24:20 – What is IBM using Quantum Computing for today?

* 00:26:00 – Wrap Up.

Scott’s Bio

Scott Crowder is currently Chief Technical Officer and Vice President, Quantum Computing, Technical Strategy and Transformation for IBM Systems. His responsibilities include leading the commercialization effort for quantum computers, driving the strategic direction across the hardware and software-defined systems portfolio, leading the agile and Design Thinking transformation, and accelerating innovation within development through special projects.

Copyright (c) 2016-2018 Storage Unpacked.  Post #Z5RL.  No reproduction or re-use without permission.

Image by UCL Mathematical and Physical Sciences from London, UK (Quantum refrigerator at UCL) [CC BY 2.0 (https://creativecommons.org/licenses/by/2.0)], via Wikimedia Commons