Technology:Unpacked

We’ve started to see the consolidation of storage vendors as some startups and long-term players in the market get acquired. Is the reason for this buying spree one of positive growth, or a defensive position to maintain survival? Chris and Martin discuss the issues and the vendors doing the buying.

Who’s been buying? Violin Systems acquired part of X-IO (specifically the ICE) products as that company changed focus to their edge device (Axellio). DDN acquired Tintri and Nexenta. StorCentric, founded from Drobo and Nexsan has acquired Retrospect and Vexata.

Are we seeing a move to a new style of holding company? Maybe these vendors are keeping assets warm, either as part of OEM agreements or for future acquisition.

Elapsed Time: 00:34;14

Timeline

* 00:00:00 – Intros 00:01:28 – Vendor acquisitions – there’s a lot of them!* 00:02:09 – Why are acquisitions happening? * 00:04:20 – Violin Systems acquires (part of) X-IO * 00:08:53 – DDN – Tintri, Nexenta * 00:13:50 – StorCentric – Drobo, Nexsan, Retrospect and… Vexata? * 00:18:16 – Building a portfolio or standalone products * 00:20:45 – Or perhaps a new style of VC? * 00:24:09 – Are we in a storage consolidation phase? * 00:26:05 – What is the long-term strategy for these companies? * 00:30:47 – Are there any really bad products? Or just unlucky? * 00:32:47 – Wrap Up

Related Podcasts & Blogs

* #64 – Success & Failure in Storage Startup Land* Garbage Collection – All-flash Market Consolidation

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #AEFA.

In this episode, Chris talks to Brian Schwarz, VP of Product Management for FlashBlade and ObjectEngine at Pure Storage.  ObjectEngine is a scale-out de-duplication engine that efficiently writes data to either FlashBlade or public cloud object stores.  The solution developed from the acquisition of StorReduce in 2018.

ObjectEngine was conceived when Pure Storage observed customers using FlashBlade for backup data.  The FlashBlade platform was originally developed for high-performance file-based applications like analytics.  De-duplication wasn’t integrated natively as an initial design decision.  Combining ObjectEngine with FlashBlade enables space saving ratios of around 8:1 or greater.

You can learn more about FlashBlade, ObjectEngine and other Pure technologies at Pure Accelerate.  The event will be held in Austin this year, between 15-18 September 2019.  You can follow Brian online at https://twitter.com/theschwarzbwthu (which we think is a Star Wars reference).

Elapsed Time: 00:15:56

Timeline

* 00:00:00 – Intros

* 00:00:45 – StorReduce was acquired 3Q2018

* 00:01:30 – Where was the need for ObjectEngine?

* 00:03:45 – FlashBlade was designed for low impact of CPU & Memory

* 00:04:25 – How is ObjectEngine packaged and sold?

* 00:07:40 – Has ObjectEngine driven new or existing customer engagement?

* 00:09:17 – ObjectEngine is S3 object API on the input and output

* 00:10:20 – How will ObjectEngine look in the public cloud?

* 00:13:20 – What is the impact of encryption with ObjectEngine?

* 00:14:55 – Pure Accelerate 2019 – This year in Austin

Related Podcasts & Blogs

* #51 – Pure Accelerate Pregame

* Soundbytes #008: FlashBlade 2.0 with Rob Lee at Pure Accelerate

* Pure Storage ObjectEngine for Flash-based Backup

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode #3193.

In this week’s episode, Chris talks to W. Curtis Preston.  Curtis is a long-time and well-known industry expert in the backup area and now Chief Technologist at Druva.  Data protection in a multi-cloud world introduces new challenges compared to traditional on-premises backup.  As a result, Druva has developed a cloud-native platform that protects on-premises, cloud, endpoint and SaaS applications.

What does cloud-native actually mean?  Chris and Curtis discuss the benefits of using native AWS public cloud services like S3, DynamoDB, RDS and EC2 instances.  Compared to on-premises backup, where hardware is procured to meet high watermark peaks, public cloud data protection can be scaled on demand.  This allows Druva to support customers with a cost model more closely aligned to actual usage.

What are the challenges of running purely from public cloud?  Clearly there are issues with network bandwidth and getting that first backup in place.  Druva uses Snowball Edge devices to both seed first backup and will support disaster recovery via Snowball devices in the future.

To learn more about Druva, visit www.druva.com.  To follow Curtis on Twitter, follow @wcpreston or log onto Curtis’ data protection portal, www.backupcentral.com where you can find Curtis’ new podcast.

Elapsed Time: 00:38:50

Timeline

* 00:00:00 – Intros

* 00:01:30 – Understanding Cloud-native data protection

* 00:02:09 – What data protection offers does Druva offer?

* 00:04:00 – Don’t forget to backup your SaaS data!

* 00:04:50 – Lift and shift or build backup natively for the cloud

* 00:06:40 – Dynamic scaling of AWS resources like EC2

* 00:08:28 – DynamoDB, RDS, S3, EC2 all Druva cloud components

* 00:09:30 – Charging is based on deduplicated backup data stored.

* 00:10:22 – Using Snowball Edge for “first backup” transfer

* 00:11:20 – Is there an issue running in a single Druva account?

* 00:13:20 – How does customer geography affect data backup?

* 00:14:58 – Druva create a blueprinted build to support new AWS regions

* 00:16:37 – Who pays for bandwidth and egress charges?

* 00:19:00 – Giant servers, giant tape libraries, giant networks!

* 00:20:20 – Druva does source-side deduplication to reduce network traffic

* 00:22:55 – How are multiple cloud providers supported?

* 00:24:30 – How are full server and entire data centre restores achieved?

* 00:27:00 – Druva performs “surgery” on backup images to convert to AMI

* 00:29:36 – DR is the perfect workload for the public cloud

* 00:32:00 – Endpoint analytics are already in place – e.g. ransomware

* 00:34:00 – With SaaS, the charging process is already in place for new features

* 00:36:00 – Approaching a $100m run rate

* 00:37:30 – Wrap Up

Related Podcasts & Blogs

* Talking Cloud Data Protection with Jaspreet Singh

* Cloud Field Day 3 Preview: Druva

* The Three Facets of Backup

Curtis’ Bio

W. Curtis Preston, also known as Mr. Backup.

IDC recently released their latest quarterly storage sales figures.  The data shows, yet again, that IBM sales continue to decline.  In this week’s podcast, Chris and Martin discuss the state of IBM’s storage business.  Is it time for IBM to quit?

IBM has an embarrassment of riches in storage software and hardware (or a nice portfolio as Martin puts it).  Many of these solutions have evolved from other technology, like SVC and XIV.  With the acquisition of Red Hat, IBM customers will have even more storage choice.  Does this mean more flexibility or confusion?

Re-using technology is not a bad thing (although it’s worth checking out the wiring in the image of the A9000 on this post.  However, with so many competing solutions, IBM could be in danger of confusing customers and failing to capitalise on the opportunities of integration with other platforms.

For more details on the latest IDC figures, see this post.  The graph showing IBM sales is included here for reference.

Elapsed Time: 00:26:48

Timeline

* 00:00:00 – Intros

* 00:00:50 – IDC storage data shows IBM in revenue decline

* 00:02:30 – Is IBM moving or losing their storage customers?

* 00:04:00 – What is IBM’s storage portfolio?

* 00:06:02 – What could be rationalised in the portfolio? DS8000?

* 00:08:49 – Does IBM have an issue with the marketing message?

* 00:10:00 – Could IBM have moved a lot of revenue to SDS?

* 00:12:30 – ESS back from the dead!

* 00:13:20 – But Red Hat (an IBM company) also has storage products!

* 00:16:54 – Spectrum Scale supports containers – who knew?

* 00:18:20 – Frictionless adoption – is that a problem for IBM?

* 00:20:20 – Is the storage hardware sales metric still fit for purpose?

* 00:21:30 – What do we recommend for IBM?

* 00:24:00 – Simplify!  Oh and don’t forget about tape.

* 00;25:00 – A “capacity under licence” measure would be better.

* 00:26:00 – Wrap Up

Related Podcasts & Blogs

* IDC 1Q2019 storage data shows a tough market for appliance vendors

* #43 – All-flash Market Review 2018 with Chris Mellor

* IBM Buys Red Hat

* IBM Reuses Existing Tech in Latest FlashSystem Release

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 12B0.

 

In this second episode on VAST Data, Chris and Martin continue the discussion with Howard Marks.  You can find the previous episode at #105 – Introduction to VAST Data (Part I).  This time, the conversation continues where the discussion left off, with Howard finishing the explanation of wide striping.

To explain exactly how data is accessed on the platform, Howard introduces the concept of v-trees.  These are like b-trees but flatter and wider.  The v-tree is used to hold both metadata and data.  One interesting aspect of the discussion is in understanding exactly how Optane and QLC are used.  As data is written, Optane acts as a write cache but doesn’t need to immediately be flushed, as would happen in traditional systems.  Instead, the VAST platform can take time to destage to QLC when appropriate, adding time for data reduction tasks to take place.

Howard wraps up the conversation with some detail on use cases, which at this stage are focused on “data intensive” applications.  In time this will expand to meet traditional enterprises too.  Roadmap items include SMB, snapshots and replication.

You can find more information the VAST Data platform at www.vastdata.com.  Follow VAST Data and Howard on Twitter.

Elapsed Time: 00:21:18

Timeline

* 00:00:00 – Intros

* 00:00:43 – Continuing the data layout discussion

* 00:02:58 – V-Trees – metadata and data

* 00:05:00 – Data placement based on type, gaming the system?

* 00:08:40 – What performance can customers expect?

* 00:11:00 – Does it make sense to add block protocols to VAST Data?

* 00:12:30 – What are the deployment models, SDS, appliance?

* 00:15:00 – Still evolutionary, Martin? Revolutionary pricing?

* 00:16:48 – Now we know the technology, what are the use cases?

* 00:19:30 – Wrap up

Related Blog Posts & Podcasts

*  #105 – Introduction to VAST Data (Part I) with Howard Marks

* VAST Data launches with new scale-out storage platform

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode FC05.

This week, Chris and Martin talk to Howard Marks, Chief Storyteller at VAST Data.  You may know Howard as an independent analyst and author for a range of online publications.  Howard recently joined VAST to help explain and promote understanding of their data platform architecture.

The VAST Data platform uses three main technologies that have only recently emerged onto the market.  QLC NAND flash provides long-term, cheap and fast permanent storage.  3D-XPoint (branded as Intel Optane) is used to store metadata and new data before it is committed to flash.  NVMe over Fabrics provides the connectivity between stateless VAST front-end servers and JBOF disk shelves.

The architecture has some very subtle differentiated points that allow the solution to be scale-out and highly efficient.  The server components are stateless because metadata isn’t cached locally. That removes issues of cache coherency and keeping all metadata synchronised.  3D-XPoint allows data to be written as huge stripes with as little as 3% overhead on large systems.

If you want to learn more about the VAST platform, check out https://www.vastdata.com, read our blog on the VAST technology or visit the Tech Field Day website where you’ll find more in-depth videos from the founders of the company.

Elapsed Time: 00:30:49

Timeline

* 00:00:00 – Intros

* 00:02:00 – Who is Howard Marks?

* 00:02:30 – Who are VAST Data?

* 00:04:00 – Do we need hyper performance or good enough?

* 00:07:30 – Three technologies – QLC NAND & Optane

* 00:09:30 – Intelligent JBOFs

* 00:11:50 – Take a breather Howard!  Let’s review!

* 00:13:30 – Server components are stateless containers

* 00:14:50 – Shared nothing? No DASE – Shared Everything

* 00:17:40 – Why does Persistent Memory allow scale-out?

* 00:19:34 – Wide stripe optimisation with erasure coding

* 00:21:00 – Optimised deduplication with similarity hashing

* 00:24:00 – Wide stripes with sequential I/O improves endurance for flash

* 00:26:00 – It’s a log-based file system (not WAFL)

* 00:29:00 – 10 year guarantee on QLC drives

Transcript

 

Related Podcasts & Blogs

* VAST Data launches new scale-out storage platform

* QLC NAND – What can we expect from the technology?

* What are TBW & DWPD?

Howard’s Bio

Howard Marks is VAST Data’s Technologist Extraordinary and Plenipotentiary helping customers realize the advantages of Universal Storage. Before joining VAST Howard spent 40 years as an independent consultant and storage industry analyst at DeepStorage. He is a frequent, and highly rated speaker at industry events and Storage Field Day delegate.

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode FC04.

This week, Chris talks to Paul Stringfellow, Technical Director at Gardner Systems, about the process of creating a data management strategy.  As we adopt additional, sometimes disparate services from SaaS and IaaS vendors, increasingly businesses are seeing their data being dispersed across multiple platforms.  With such a valuable asset at their fingertips, how do businesses go about building a strategy for storing, managing and securing their information?

The conversation starts with a look at three layers – physical infrastructure, smart storage management and data management.  Storage platforms are pretty fully functional these days, so features like APIs, protocol support and data services are “table stakes”.  Vendors need to provide these as standard and most generally do.

What about the public cloud?  How as this changed the perception of data management?  As a storage medium, cloud is a great resource, but comes with a different charging model.  Businesses need to be sure that data in the cloud is being stored for some current or future value.  Otherwise, recurring costs can be expensive.

What are the steps to building a data strategy?  The initial focus is identification, a challenge in modern businesses, where data can be spread across sources including on-premises and cloud infrastructure.  Owners need to be identified and requirements collected.  This needs to be used as the basis for technology selection.  This includes looking at the wider aspects of application architecture.

You can find Paul on the Twitters at @techstringy and at www.techstringy.com.  His podcast (Tech Interviews) is online at podcast.techstringy.com and SoundCloud.

Elapsed Time: 00:41:22

Timeline

* 00:00:00 – Intros

* 00:02:00 – Three layers, physical infrastructure, storage and data management

* 00:02:40 – Is storage hardware good enough these days?

* 00:04:30 – Table stakes – protocols, APIs, data services

* 00:06:34 – What is the impact on (public) cloud in a data strategy?

* 00:09:00 – Is public cloud being used tactically for data management?

* 00:10:30 – Does public cloud change behaviour in retaining data?

* 00:13:30 – Cloud changes the storage cost focus from every 5 years to every 4 weeks

* 00:18:00 – What are the steps to build data management strategy? Start with Audit

* 00:20:40 – Identifying data is harder than ever, with on-prem, cloud and SaaS services

* 00:25:30 – Who owns the responsibility of managing business data and policies?

* 00:28:30 – Data is your of information

* 00:30:16 – What components go into a data platform?

* 00:35:39 – CIOs/CTOs need to look at application architecture as part of a data strategy

* 00:38:45 – Everybody loves shiny

* 00:40:00 – Wrap Up

Related Podcasts & Blogs

* #103 – Data Management and DataOps with Hitachi Vantara (Sponsored)

* #81 – Storage or Data Asset Management?

* Software Defining Your Storage

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 6773.

This week Chris speaks to Jonathan Martin (CMO) and John Magee (VP, Portfolio Marketing) at Hitachi Vantara.  This episode was recorded live onsite at the new Hitachi Vantara offices in Santa Clara.

As data becomes ever more valuable to organisations, the process of building data pipelines will continue to be a time and resource intensive task.  In order to keep up with demand, Hitachi Vantara believes that businesses will need to implement automated processes that deal with the analytics pipeline – a concept called DataOps.

DataOps is a methodology rather than any specific product.  It defines automated processes that both improve the quality and reduce the cycle time in analysing and exploiting large volumes of data.  In a similar process to the way in which DevOps allowed continuous development and deployment of applications, DataOps looks to bring continuous improvement to data and analytics.

Surely, there will be challenges?  Governance still needs to exist around how data is used within an organisation.  There will still be a need for data owners and good practice in protecting, curating and managing data.

Ultimately, as Jonathan explains, we’re creating so much data that centralising it will be impossible to achieve.  Instead, business will need to develop distributed models for data management and that is likely to be a major challenge for the foreseeable future.

You can learn more about DataOps either from following Hitachi Vantara at www.hitachivantara.com or reading the Hitachi blogs.  Alternatively, join Hitachi Vantara for Hitachi .NEXT 2019 in Las Vegas in October.

Elapsed Time: 00:29:37

Timeline

* 00:00:00 – Intro

* 00:01:00 – Comparing infrastructure versus data management companies

* 00:02:30 – Infrastructure still forms the foundation of Hitachi products & solutions

* 00:03:30 – Modern data challenges include diverse sources of data

* 00:06:00 – How do customers deal with scale in designing data pipelines?

* 00:06:48 – What is DataOps?

* 00:10:32 – How will new data sources be curated and governed?

* 00:12:00 – Data Management for the AI era

* 00:13:40 – Could automation all data to be misused?

* 00:15:15 – What technology is in place to enable DataOps?

* 00:20:30 – CarFax – bringing multiple data sources together with Hitachi technology

* 00:22:27 – How are customers looking at hybrid cloud?

* 00:24:30 – Designing a distributed data model will be an ongoing data challenge

* 00:27:00 – DataOps – learn more at Hitachi .NEXT

* 00:28:50 – Wrap Up

Related Podcasts & Blogs

* #72 – Hitachi Vantara Looking Forward with Shawn Rosemarin

* #67 – Hitachi Vantara ne Year On with John Magee

* #70 – Pentaho Data Aggregation with Arik Pelkey

* #68 – Intelligent Object Storage With Scott Baker

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 1CD0.

This week Martin, Chris Evans and Chris Mellor discuss Dell EMC’s plans for merging the three current midrange storage platforms marketed by the company.  This is a follow-up to the article Chris recently posted on https://blockandfiles.com that discusses Project Trident.

There are a lot of moving parts to consider when merging storage platforms.  Customers have spent time and effort developing processes and scripting to automate common processes.  How will these be supported?  Will the new platform appear to take on the persona of the previous three?  Should customers (as Martin believes) be forced to make a clean break and throw away old scripts?

In the file and object space, the team think Islion will continue to have a prominent place in Dell EMC’s portfolio.  What about the high-end could Dell EMC acquire or build a high-performance scale-out file system?

The conversation turns next to acquisitive DDN, which has recently acquired Nexenta.  DDN also bought out the assets of Tintri.  Is DDN looking to move to the enterprise market?

Finally, the team consider HPE’s acquisition of Cray Inc.  This now gives HPE a range of storage platforms for HPC, but will the strategy turn out to be successful?

Elapsed Time: 00:31:54

Timeline

* 00:00:00 – Intros

* 00:01:00 – Dell EMC – a plethora of midrange storage arrays

* 00:02:20 – Merging will just create a new CLARiiON?

* 00:02:40 – What is the result of 3:1 merging?

* 00:03:49 – Pain-free migration, how does this work with scripts?

* 00:06:20 – Maybe vendors should be forced to dump old scripting

* 00:07:30 – Which platform will be the favoured one for Midrange.NEXT?

* 00:09:14 – Unity has the biggest user base – so will be the survivor?

* 00:11:20 – Which platforms are the most feature-rich?

* 00:12:30 – How should engineering teams be restructured?

* 00:13:57 – Does NetApp history show how not to merge platforms together?

* 00:16:00 – What is Trident?

* 00:18:40 – As a customer what assurances would you be asking your vendor for?

* 00:19:36 – Are other vendors looking to steal customers as part of the consolidation?

* 00:21:00 – Isilon provides scalable services for scale-out file requirements

* 00:23:12 – Do we have too many midrange storage vendors?  DDN acquiring?

* 00:25:19 – Chris is charmed by Tarkan! Or is it just storage CEOs?

* 00:27:17 – HPE buys Cray – what’s the reason behind this?

* 00:31:00 – Wrap Up….

Related Podcasts & Blogs

* #90 – Dell EMC’s Enterprise NVMe Strategy with Vince Westin

* #99 – Dell Technologies World 2019 in Review

* What Next for XtremIO?

Copyright (c) 2016-2019 Storage Unpacked.  No reproduction or re-use without permission. Podcast Episode 0B2C.

In this podcast episode, Chris talks to Brian Biles (Chief Product Officer and co-founder) and Tim Page (CEO) from Datrium about the announcement of Automatrix.  The Datrium Automatrix platform implements five important components needed to deliver a consistent approach to application mobility.  These are primary storage, backup, disaster recovery, encryption and data mobility.

Automatrix brings together existing products that include DVX and Cloud DVX with the general availability of ControlShift (previously Project CloudShift).  ControlShift provides full automation of the disaster recovery failover and failback process, currently between on-premises DVX instances and by the end of 2019, to and from VMware on AWS.

In this discussion, Brian explains the importance of combining the five components into a holistic solution.  Without the ability to implement a consistent data plane across clouds, implementing application mobility becomes complex and expensive.  Automatrix provides the data plane, onto which virtualisation can be layered, either in the form of VMware or through containers.  Today container support means Docker, but will be extended to Kubernetes in the future.

Find out more about Datrium and Automatrix at https://www.datrium.com.

Elapsed Time: 00:35:50

Timeline

* 00:00:00 – Intros

* 00:01:00 – Background on Datrium – DVX

* 00:04:10 – What is Cloud DVX?

* 00:05:10 – What is ControlShift

* 00:06:55 – How much of a problem was automated DR for customers?

* 00:08:30 – DR is complex, with runbooks and compliance requirements

* 00:11:30 – How do you know where you backed up an application 6 months ago?

* 00:13:30 – Reducing DR costs represents a big saving for enterprises

* 00:15:30 – Surely primary and secondary data not the same platform is a problem?

* 00:18:00 – DR & backup/restore are pretty similar – just differing on scale

* 00:21:50 – What are the new announcements from Datrium?

* 00:23:00 – Five components, primary storage, backup, encryption, mobility, workflow orchestration

* 00:25:55 – Insurance shouldn’t be more expensive than buying the product itself

* 00:26:50 – ControlShift is now GA for on-prem to on-prem workloads

* 00:27:40 – ControlShift offers much more than DR – optimisation, consolidation

* 00:31:00 – Virtualisation layers are standardising – VMware & Kubernetes

* 00:31:54 – What’s the future? VMware on AWS support, primary storage on AWS, Azure, Kubernetes

Transcript

 

Brian’s Bio

Brian Biles is Chief Product Officer and co-founder of Datrium. In 2001, Brian was a founder of Data Domain, a pioneer in building large-scale data deduplication systems. He served as VP of Product Management and Business Development until the company’s acquisition by EMC in 2009, where he became VP of Product Management, Backup Recovery Systems. Previously he held leadership positions at Sun Microsystems, VA Linux and Data General. He has a BS in Computer and Information Sciences from the University of California, Santa Cruz. In 1998, Brian earned a “special thanks to” credit in the movie The Big Lebowski.

Tim’s Bio

Tim Page is the CEO of Datrium. He was previously one of the founding members of VCE, helping grow that venture to $3 billion in revenue in 7 years.