Architecting IT

3PAR Continues to be HP Storage Cornerstone

Chris M Evans — Wed, 12 Jun 2013 17:25:48 +0000

This week HP announced their new all-flash array, the 3PAR StoreServ 7450. I’ll post separately on the details of the new platform, but for now, I’d like to discuss the strategy HP are taking compared to the wider market.

As we can see from one of the slides presented yesterday, HP are firmly targeting EMC’s disparate technology portfolio. It’s true that EMC have many different technologies they have acquired over the years, however in general these products meet a range of storage needs, so we should only be considering those within the platform types. Even so, in primary storage we see 5 separate offerings. This in itself isn’t a big deal unless you have a need to work across these platforms and want seamless and consistent management across them all.

HP continue to stress the benefit of the single 3PAR StoreServ platform in its ability to range from entry level (7200 model) to the high end StoreServ 10800. The newly announced 7450 fits the “performance” requirement slot as can be seen on the second slide. This slide also highlights the merits of a single interface, single technology and so on. Personally I think this is a great benefit for a number of reasons including:

Data Mobility – I can move data in and out of my flash architecture device, using a single (built in) replication technology. So, if I have workload that needs SSD performance, I can move it between arrays reasonably simply.
Understandable Performance – the 3PAR architecture uses a dispersed architecture for storing data across as many disks or SSDs as possible. The only difference between a standard array and the all-flash models is the speed of those devices. This makes it easy for me to understand the amount of performance scaling I should expect without having to think in more detail about a new design.
Consistent Management – users can configure through a single interface and only have to go through one set of integration tasks for monitoring, alerting, security etc.

Could other vendors do this? Of course they could. It was interesting to see Hitachi was missed off HP’s competitor slide. Their new flash blades and chassis for VSP are a realistic competitor to HP and provide the same scalability, consistent management and mobility features HP are claiming for 3PAR. In addition, HP can support mainframe, where 3PAR can’t. The funny thing is, at HP Discover in Frankfurt, HP were promoting the XP24000, the VSP equivalent, including the flash blades, meaning HP still rungs two platforms of their own.

The Architect’s View

In his keynote today, Dave Donatelli claimed that flash startups don’t have the breadth of features that the 3PAR platform offers. This is somewhat disingenuous to the many companies that have created some true innovation in this new space. However this is a standard marketing approach and we shouldn’t expect anything less. We see worse from HP’s competitors. The 3PAR platform does have an architecture that works well with SSD and both hardware and software changes have been made to the 7450 to optimise performance. I’ll be posting more on this later, as the detail is what really counts for customers.

Disclaimer: HP have paid for flights and accommodation for Chris Evans to attend HP Discover. However this does not imply a requirement to blog about the event and no editorial rights to any published content is provided.

Comments are always welcome; please indicate if you work for a vendor as it’s only fair. If you have any related links of interest, please feel free to add them as a comment for consideration.

Subscribe to the newsletter! – simply follow this link and enter your basic details (email addresses not shared with any other site).

HP Discover – Las Vegas 11-13 June 2013 and Software Defined Storage

Chris M Evans — Fri, 07 Jun 2013 13:22:41 +0000

This coming week I will be in Las Vegas for HP Discover 2013. I was last here 2 years ago and HP have had a turbulent time since then, especially with their mobile strategy and the Autonomy acquisition. From my perspective, I’m interested to hear about new developments on the storage front – no doubt there will be announcements from the 3PAR (StoreServ) and StoreVirtual teams amongst others. I’m also looking forward to getting more information on Project Moonshot, billed as a software defined server. I’m not really clear on the value proposition for Moonshot and having yet another “software defined” moniker doesn’t help.

Speaking of software defined, I did a video with HP a couple of weeks ago on software defined storage. Although we still have no clear definition, Kate Davis and I discussed some of the ideas and issues around what software defined storage means. A link to the video is listed here.

HP Tech Talk – Software Defined Storage

Based on my poll earlier this year, many people see SDS as being based on “white-box” x86 servers. I’m sure that will please Intel, but fundamentally from a hardware perspective I think we need more than that in certain instances. For example, exploiting flash fully still requires some design around the hardware; building in fault tolerance & availability (e.g. hot-swapping components) requires a clustered design – this isn’t simply white-box x86 implementation. I do agree with the policy-based management piece though and that ultimately for me, along with QoS is what I see SDS delivering. Historically, storage arrays have delivered “best efforts” responses, with lightly loaded arrays outperforming a fully loaded configuration; we all can point to examples where end users who were first on a new array would start complaining once the box was fully configured to other hosts. What we do need is more predictable response, even if that means under-utilising hardware initially. It also provides the basis for differentiated service; something providers can charge for.

If you’re in Vegas drop me a note, otherwise I’ll be based around the bloggers area. In the meantime, have a look at the following link that covers potential new storage announcements. I’ve also included a link to Calvin Zito‘s list of attending bloggers.

Roll Call: Bloggers Coming to HP Discover

Countdown to HP Discover: Storage built for the next era

Interested in learning more about some of the hottest storage trends in 2013 such as flash and software-defined storage? Join us on June 11th at 8:30 AM Pacific LIVE from HP Discover in Las Vegas to learn how HP Converged Storage is transforming the industry!

RSVP today! http://hpstorage.me/12aj51C

Comments are always welcome; please indicate if you work for a vendor as it’s only fair. If you have any related links of interest, please feel free to add them as a comment for consideration.

Subscribe to the newsletter! – simply follow this link and enter your basic details (email addresses not shared with any other site).

Will Connected Data be Good for Drobo?

Chris M Evans — Wed, 05 Jun 2013 12:19:15 +0000

Word on the street says Drobo are planning a merger with Connected Data, the new company at the heart of the transporter product. Drobo, formerly Data Robotics Inc, makes a range of “storage robots”, devices that need little or no user management and come in a range of consumer and SMB offerings. Connected Data recently released the Transporter, a desktop storage device that provides peer-to-peer synchronisation between devices without storing data in “the cloud”. Both companies were founded by Geoff Barrall, a long-time storage innovator, who also founded BluArc, acquired last year by Hitachi Data Systems.

At the outset this deal seems a little odd and appears almost to be a reverse takeover, with the smaller Connected Data swallowing up Drobo. There are no details on Drobo’s website and finding any comment on the Connected Data site proved difficult in the extreme. So, at this stage we have little information to go on, other than a number of posts from Chris Mellor at The Register.

Digging a little deeper, there are a number of issues Drobo are struggling with. Many see their technology as overpriced and under-featured (if that’s a valid expression) and one big gap is native replication within the devices. For example, the SMB 1200i/800i needs facilities to replicate between devices; there’s also a huge opportunity missing where Drobo could allow their “pro-sumer” products to replicate to data centre based 1200i/800i devices as a central archive store. Only the 800fs model has native replication and this is a pure NAS (CIFS & AFP only) offering.

Connected Data have file-based replication software, but more usefully, IP and knowledge in multi-device synchronisation and data consistency. There’s a perfect opportunity for that technology to come back to Drobo and be incorporated into their existing and future product set. From Connected Data’s point of view, they gain an instant customer market, access to other hardware into which to place their software, many of those being consumer devices. A potential win-win.

The Architect’s View

A Drobo/CD merger could be good news for both parties. I like Drobo, both as a company and the concept of their technology and I would like to see them continue to do well. However the consumer/SMB storage market is now massively competitive and Drobo’s lack of features and high-end cost must be hurting. Hopefully, a rejuvenated Barrall can add a touch of creative spark that Drobo badly needs.

Managing Microcode Upgrades

Chris M Evans — Mon, 03 Jun 2013 14:15:50 +0000

Blog reader Chris Saunderson contacted me some time ago and asked the question on how often array microcode should be upgraded. Well, better late than never Chris, I’ve crafted the following blog post to discuss the subject.

Chris posed the question in relation to storage microcode. On many (certainly enterprise) storage arrays, the control functions are managed in microcode, which is still essentially the software that is running the array. Looking back to older models, that code could be running on various components and in many cases is fault tolerant, running independently on multiple controllers within the device.

Harking back to my mainframe days, I used to be a systems programmer and did a lot of system upgrades, installation and maintenance. Although it was tedious, we typically followed a process of evaluating the impact of every software release. If a release had no intrinsic benefit, then we wouldn’t implement it. However once every three or six months we would do a rollup of all outstanding maintenance and apply it simply to keep up to date. There are a number of reasons for this:

Applying updates brought in new features which may prove useful in the future
Applying scheduled updates implemented bug fixes we weren’t aware of, but may cause problems in the future. I’d rather have a bug fix in place than find it by hitting the bug.
Applying updates on a regular (if not frequent) schedule reduced the risk of hitting “showstopper” problems that required rapid testing and implementation, potentially impacting service.

Unfortunately things in the storage arena are a little more complicated than in mainframe days. Shared storage can have tens or hundreds of connected servers. An upgrade to storage array microcode could have an impact on storage networking components, host driver & firmware levels and management applications, so upgrades are not a thing to take on lightly.

Here’s how I would work it.

Work with your storage array vendor to determine their upgrade cycle. Do they do annual major upgrades? Do they do point releases or patch releases on a more frequent level?
Plan a fixed number of maintenance slots per year, say two (6 monthly) or four (quarterly). Agree that some may need server downtime. Keep the slots, even if they are not used, as it gets users, change control and others used to the idea of regular schedule maintenance slots. Match the slots to the vendor upgrade cycle, but with a time lag that allows testing and allows for the code to be out in the field and bug tested by other braver souls.
Don’t upgrade all arrays at the same time. Start with test/development boxes and roll out to the most mission critical environments last.
Only allow bug fixes to be applied where they address a direct issue being experienced in your environment. Otherwise wait for the patches to appear in a major release.
Keep driver/firmware up to date as much as possible on the host. Maintain an inventory of these versions so microcode can be validated against them.

There’s a lot of effort required to keep shared storage environments current. So, regular updates and maintenance are essential across the whole stack and should be carried out continually, to both reduce the risk of failure/downtime and to reduce the number of “forced” upgrades that can inevitably happen if environments aren’t regularly maintained.

I’d be interested to hear what other users do and their thoughts around staying current.

ViPR – Frankenstorage Revisited

Chris M Evans — Thu, 09 May 2013 12:03:25 +0000

Cast your minds back to 2009. At that time, Chuck Hollis wrote a blog post that quoted the following definition:

“Frankenstorage appears to be a new twist on [the Frankenstein] idea — storage arrays, assembled from various parts from multiple vendors, brought to life by the magic of powerpoints and press releases.”

The crux of the blog post referencing this lovely term was storage virtualisation. Apparently in 2009, storage virtualisation was bad. Why would anyone want to take the opportunity to use a consistent access method for bringing disparate data sources together under one common standard protocol? Madness.

Wind forward to 2013 and we have a very different landscape. EMC VMAX now supports external storage; VPLEX effectively virtualises other storage platforms, although we have to call it storage federation. And now we have EMC ViPR, which apparently introduces storage virtualisation too – albeit in a new form only covering the management aspect of storage infrastructure.

Project Bourne

ViPR is apparently the crystallisation of Project Bourne, a (not too secret) piece of work from EMC’s Advanced Storage Division. The marketing message tells us that we need to separate the control plane from the data plane in storage (as has been achieved with Software Defined Networking) and ViPR delivers that for us, with all the perceived benefits. I have a feeling the reality is probably something different.

Although EMC continues to be the leader in storage array sales (see my recent Gartner post covering this), they are seeing strong growth from their competition. One of the major issues with EMC’s VMAX and VNX platforms is storage management at scale. EMC’s SRM tools have never fully delivered and been complex and cumbersome to deploy and operate. As a result, many companies still provision and do migrations via scripts. This is a tedious and error-prone process, requiring significant understanding of the underlying platform to ensure workloads are balanced efficiently. EMC’s competitors have done a better job at managing the storage management issues; Hitachi have invested heavily in their software and can do tasks like migration without outages using UVM and HAM. HP 3Par is a management dream compared to provisioning storage on VMAX. Then there is the breed of new devices and startups (e.g. SolidFire) that have focused on simplified management and native API integration. Note the use of the word “native” here, not some new artificial wrapper like VMAX Cloud Edition.

iWave

So EMC needed to do something to fix this problem; they acquire iWave Software, a startup focusing on storage automation software tools. Their Storage Automator software takes some of the manual work out of storage provisioning and provides a framework into which storage arrays can be placed. Each array can be attributed to more generic technology pools (e.g. gold/silver/bronze tiers) from which storage is provisioned to hosts, including all the fabric zoning and LUN masking needed. Storage Automator provides a number of benefits; it enables lesser skilled staff to do storage provisioning within the controls of the framework established by storage architects. It implements automation around provisioning out of hours and it abstracts the view of the array to more generic tiering constructs.

I first saw iWave in March 2012 and was impressed with the product. There were a few rough edges and platform support was limited, but as a tool to reduce the time and effort of provisioning, it did a good job. What it did not do was storage virtualisation. LUNs were presented to the host directly from the managed devices with no other physical abstraction. There were no features for load balancing across pools, multi-tenancy (other than allowing portions of the pools to be provisioned by certain users), transparent migration, or any of the other features that would commonly be expected from a storage virtualisation platform.

ViPR

The Homer

EMC’s acquisition of iWave in January 2013 (although not formally announced) provided them the ability to make good on some of the provisioning shortcomings, however that wouldn’t be enough for Jeremy and the boys in marketing. What they wanted was a stronger message and what better vehicle than Software Defined Storage. Provisioning automation gets positioned as the control plane, which leaves the data plane issue to solve. This is where Frankenstorage comes in again. iWave has been combined with some HDFS/NFS presentation software layer to create this new product that is “assembled from various parts from multiple acquisitions [vendors], brought to life by the magic of powerpoints and press releases”.

This intention to merge what at the outset appears to be two distinct classes of functionality is probably the most confusing. At this stage I can’t see how ViPR could truly be called storage virtualisation unless they incorporate VPLEX into the mix and then we’re really into Frankenstorage solutions. At some point you have to stop bolting bits on and accept legacy technology is past it’s sell by date.

The Architect’s View

As a $50 billon company, I’d expect more out of EMC than repackaging of seemingly unconnected components. The marketing message seems to be overtaking the technology to a point where they are believing their own hype. If you’re looking at ViPR, question EMC on what the evolution of this platform will be; question exactly how multi-tenancy, rich APIs actually work; above all, measure the value proposition of their solution against the rest of the market, because the storage landscape is much bigger (and in many cases much better) than just EMC.

Gartner ECB (External Controller-Based Disk) – 4Q2012

Chris M Evans — Tue, 26 Mar 2013 07:31:42 +0000

Figure 1 – Gartner EBC – All Vendors

The latest Gartner ECB survey is out and as discussed in my last post, I’ve updated my data to reflect the latest set of figures. The vendors all have ups and downs each quarter so this data is particularly interesting as it covers year end 2012 and allows four years of comparisons from 2009 through to 2012.

As expected, in absolute terms, EMC continue to power ahead, as is indicated in the “all vendors” graph on the left (figure 1). What’s more interesting is the obvious sales cycle the vendors go through. For EMC, quarter 4 is the important one as that’s where they receive the biggest uptick. IBM have a curious but consistent cycle that declines heavily but also rises heavily too, and again is remarkably consistent. The other vendors also have similar cycles (which no doubt fit their own financial year ends). Hitachi’s is interesting because for them, the jump from Q3 to Q4 has typically been high, with other period seeing a drop. Over this quarter that expected drop hasn’t been seen, but rather they saw a slight rise. Does this demonstrate an improved growth business?

Figure 2 – Gartner ECB 3 Year Relati

Oracle’s graph continues to look dire, although this quarter they saw only a slight drop on the previous one. The polynomial curve mapped against the figures starts to show a slight turn towards bottoming out rather than a continual slide. NetApp’s figures show a flattening out of their growth compared to great performance from 2009 onwards. Overall, only EMC and Hitachi continue to show straight line growth, although I should stress that this is a small sample of data and it can be interpreted many ways.

Finally, the relative 3 year growth graph, comparing figures from 2009 onwards can now be presented with a full year for 2012. This is shown in figure 2. EMC continue to grow consistently strongly. NetApp have declined slightly and their growth has slowed to match that of EMC. Hitachi continues to be number 3 over the period with consistent growth figures. The remainder have either declined slightly, or in the case of Oracle, declined significantly. As previously mentioned, Oracle’s decline appears to be flattening out.

The Architect’s View

These figures can be viewed in many ways, however there is no doubting (or denying) the continued strength of EMC’s business. NetApp seem to have slowed slightly and I wonder how much of that is a reflection on the crossroads many customers find themselves at with Data ONTAP and the migration to cluster-mode. Hitachi continue to demonstrate good figures and that reflects what they have presented at their Influencers’ Days and what I’ve heard on the ground from customers and other vendors. As for the others, their business remains flat and they need some injection of ideas or technology to get thing moving.

If these trends continue, then we’re going to see an increasingly three-horse race in storage. That’s not surprising if we look at Gartner’s new Magic Quadrant for General Purpose Disk Arrays, which as expected rates EMC, NetApp and Hitachi as the top three visonaries and deliverers – as it would using this source data which is from Gartner itself.

Death of The PC (Or Mine At Least)

Chris M Evans — Sun, 24 Mar 2013 18:07:08 +0000

Today marks the first time in probably 20 years that none of the computing devices in my home are PCs. The homebrew PC tower unit used until recently by Son#2 is being shipped out after being stripped of anything useful. He now uses a Raspberry Pi and screen to RDP into a virtual desktop in the lab.

Looking around at how our computing at home has changed, it has been an interesting evolution:

Gaming has become a PS3, iPod and iPad pastime, with the PC relegated to playing Age of Empires on an occasional basis.
Content creation is a laptop thing with three in the household.
Media consumption has moved to iPads & iPods/iPhones for music, video and books.

Although I still retain my desktop iMac, I can see a day when even that will be redundant and I replace it with a large screen (or two) connected to my MacBook.

I have no idea whether my experiences are typical, but I suspect they might be. The PC as a platform (in the form of a tower or desktop) has pretty much had its day. Today’s laptops or MacBooks can offer all the performance and graphics capabilities of all but the high-end PC devices. Tablets and iPods/iPhones have effectively owned the media consumption space. No-one surely watches anything on a desktop PC any more.

As the PC hardware platform has been superseded, so will the traditional Windows operating system. iOS on Apple devices shows us that the need to understand and use the O/S at the base level is no longer necessary. What does that mean for Windows? Well, from a server perspective, it will continue to live. From a consumer perspective, it will have to mutate into the form we see on Windows 8 mobile and tablets, where the user is obfuscated from the need to understand what’s going on under the covers. This means become much more slick on patching, platform management and so on. However I think it could be a step too far for Microsoft to achieve.

Getting back to the original discussion, I look back with fond memories of the PC as well as some frustrating times when things didn’t work as I wanted. On balance, the PC was a great platform but the future of end user computing will no doubt look very different from the PC of 1981.

Google’s Hollow Cloud Storage Offering

Chris M Evans — Mon, 11 Mar 2013 08:19:52 +0000

Reading this recent article on ZDNet I was struck by the pointlessness of Google’s 1TB of user storage offering. This may seem like a bold statement, but let’s look at the figures and the offering. If you buy a Chromebook Pixel, it comes with 1TB of Google Drive storage for three years. Sadly the ZDNet article in question goes on to try and highlight how this is a great deal, if you need 1TB of space.

However, what doesn’t get discussed is the time it would take to both make use of that storage and get off it in three years’ time if you went with another storage provider.

Average UK broadband speed is now apparently 9Mb/s, although there are hotspot areas (me included) where bandwidth is available as high as 100Mb/s. I do see that speed at quiet times during the day (in fact I’ve seen it as high as 120Mb/s) but in reality at busy times it doesn’t get above 20Mb/s. It’s also worth remembering that almost all the broadband delivery in the UK (even the super-fast Virgin service) is asymmetric, so download is much faster than upload. Upload is what counts for using cloud storage, because if we can’t get it into the cloud – then we can’t download it.

If we do calculations on throughput, making simple assumptions of 100% efficiency on the line, 8 bits per byte and a 10:1 ratio on upload speed, then we see that at 9Mb/s, it would take 105 days to upload 1TB of data to Google Drive. In reality we’re never going to achieve that level of efficiency, or dedicate all our bandwidth traffic to uploading files to Google. In addition, broadband providers have limits on upload volumes too. Even if we use the full bandwidth during an eight hour day, that still means it would take almost a year to load up 1TB of content.

There are a few other things worth noting here. The terms and conditions for the free space offer seem a little confused, referencing smaller capacity values. However moving past that, we see that the 1TB offer is per Google account. Buying 10 Pixels only gets you the 1TB of storage, not 10TB.

Second there’s the question of caching. The Pixel only comes with 32GB of SSD, of which the usable capacity will be less. This means an awful lot of network traffic back and forth to access your content. At best, only 3% of your content will be on the local device. This again cuts into your upload bandwidth capacity.

Third, if you’re using a Google Drive account on non-Google devices, then all of the content will be stored locally (except any Google Docs files). So, if you share your Google account across devices, your capacity will be limited to the hard drive size of your biggest device, not the 1TB figure.

Lastly, we should talk about the cost of providing this storage. A single external 1TB hard drive can be purchased for as little as £55 ( $85) on Amazon today. Retail 3TB hard drives are as low as £92 ($138), or $46 per TB. Assuming Google even pay retail prices (and we know they won’t), then even with a mirrored configuration (RAID-1), Google have made their money back on the disk in 2 months, during which, you will have failed to upload anywhere near 1TB of content to make use of your allowance.

The Architect’s View

What seems like a good deal frequently isn’t. Retail prices for cloud storage far outstrip the cost of using local drives. Of course large data centres cost money to build and run and that adds increases Google’s cost of providing that 1TB to you. However, some people in the industry have decided that we don’t need a sledgehammer to crack a nut. The Transporter, for example lets consumers share data without requiring it being placed into the cloud. Cloud storage is good for many things, but at this stage, the financials of storing terabytes online just don’t stack up.

Manipulating the Message – The Art of Marketing

Chris M Evans — Tue, 05 Mar 2013 18:20:47 +0000

The last few weeks have seen a couple of important flash technology announcements. They are important as they bring two of the “Big 5/6″ storage vendors up to speed in a market that is already pretty well occupied. Unfortunately all of these announcements are coupled with hyperbole, figure hiding and in some cases, bare-faced lies.

Company N

Let’s look first at NetApp’s announcement of their first all-flash array, the EF540. Here are some of the bold statements NetApp make:

In a move that extends its leadership in the industry…

With the strongest and broadest flash portfolio in the industry….

…is the industry’s first flash array to combine consistent extreme performance with enterprise-class high availability, reliability, manageability, and worldwide support.

Presumably, NetApp thought by adding “wordwide support” onto that last statement, people would believe they were the first to deliver all those other features too. We all know that all-flash arrays have been available for some time. Excluding the start-ups, who may not be able to provide world-wide support, we can reference the following:

EMC all flash VNX – October 2011
EMC has shipped all-flash VMAX – May 2011
Hitachi all-flash VSP – November 2012
IBM all-flash V7000 – June 2012
3Par StoreServ All-flash – July 2012
Lefthand P4900 – February 2012

(above list updated with kind thanks to Calvin Zito)

Focusing more on the startups, I’ve discussed the players many times before. They are leading the market and purely by virtue of their startup nature may not be able to offer sales and support in every country, but if you don’t operate globally then who cares. In the NetApp announcement, reference is made to a product that will not ship for another 12 months. FlashRay seems to describe what NetApp would love to have announced now, but clearly doesn’t exist or isn’t ready (or they haven’t acquired the technology). This says that the EF540 is a a stopgap product, not a leadership position.

Company E

Then there’s EMC. Today the company announced the evolution of their server-based PCIe SSD cards, now branded as XtremSF. These appear to be rebranded Virident cards (as highlighted by Chris Mellor), with EMC software. The presented comparisons use the lower specified Fusion-IO ioDrive2 1.2TB MLC pitched against the top end XtremSF 2.2TB model. Naturally, EMC’s figures look best, but it would have been more transparent to compare the two devices with similar capacities and using the same bus architecture. For flash, more capacity means more performance, so the Fusion-IO ioDrive2 Duo has more throughput and lower latency, particularly on random writes. Take a look at the “entry level” XtremSF 550GB model and we see this is outperformed by the nearest ioDrive2 model (785GB) in every respect.

EMC also quietly dropped Thunder, their “server-side” cache product, through what appears to be lack of interest (at least Mark Twomey was honest enough to admit this). However EMC made plenty of pronouncements for that technology, but they were wrong – sometimes innovation is misconceived. NetApp need to remember that when announcing a product that will ship in 2014 (assuming it doesn’t have the Spinnaker issues).

The Architect’s View

The marketing message has reached new levels of absurdity, almost matching those we see from the music industry with a new album released. Unfortunately many customers will take the big boys’ statements on face value. Please, always remember – Caveat Emptor.

Data Storage and The Question of Trust

Chris M Evans — Mon, 04 Mar 2013 07:54:21 +0000

Nexenta, the software-only storage company received a further $24m in funding last week and had a swap around of their top execs, with former CEO Evan Powell moving to Chief Strategy Officer. Nexenta have been described by some (including themselves) as Software Defined Storage, but we’ll park that definition for now and highlight the fact that Nexenta don’t sell hardware but provide a software solution that turns servers (or virtual machines) into fully functional storage arrays, based on the core technology of Sun’s ZFS file system. What struck me as interesting about the Nexenta announcement was the volume of data their systems now manage. It’s something like 660PB with 33 customers deploying 1PB or more and 5000 customer deployments. I’m sure many people have looked at the Nexenta offering and been turned off by the fact that its software only. In reality that’s a big mistake to make. Think about what the “big 4″ storage vendors have been doing for the last few years; EMC’s VNX is a hybrid box running CLARiiON and Celerra code. Dell’s Compellent and Equallogic are basically software implementations on commodity hardware. HP commoditised all of their storage platforms bar the 3PAR and resold them on standard HP servers. IBM took SVC and developed Storwize V7000. Hitachi took their array code and made it virtual in the VSP. All the value has been in software.

Trust

However I think for many storage people, there’s been the question of trust to overcome. Mentally, there’s an assumption that a “black box” of hardware and software is inherently more reliable than running the software on your own hardware. We see it all the time; Apple make a virtue of running their code on their hardware only. Anyone who battled with early Linux distributions or Windows from 1.0 onwards knows about hardware and driver compatibility problems. Our earliest computing architectures based around the mainframe were all constructed around the premise that the vendors did software and hardware together. That makes it difficult for customers to accept that a software only solution can have value, can work on commodity hardware and probably most important, is worth paying a premium for. Perhaps this last statement more than anything is most relevant here. Martin Glassborow recently wrote about the lack of Open Source storage solutions. Software is cheap and free for many other areas, but yet again, storage is special. Users expect storage platforms to work and be 100% reliable. Where am I going with this? Perhaps to say that Nexenta have demonstrated that storage software decoupled from hardware can be reliable enough to deploy 660PB of data onto it. It paves the way for more VMs.

The Architect’s View

Whilst we struggle to quantify the definition of “Software Defined Storage” it’s clear that startups like Nexenta (and others like Nasuni) continue to bring value to software-only solutions. The big guys must be nervous as they’ve made huge amounts of margin from reselling commodity disk components over the years. This will continue to get worse for the legacy vendors as the degree of trust increases and we all become more comfortable with building our own storage arrays.

Comments are always welcome; please indicate if you work for a vendor as it’s only fair. If you have any related links of interest, please feel free to add them as a comment for consideration.

Subscribe to the newsletter! – simply follow this link and enter your basic details (email addresses not shared with any other site).

Architecting IT

3PAR Continues to be HP Storage Cornerstone

The Architect’s View

HP Discover – Las Vegas 11-13 June 2013 and Software Defined Storage

Will Connected Data be Good for Drobo?

The Architect’s View

Related Links

Managing Microcode Upgrades

ViPR – Frankenstorage Revisited

Project Bourne

iWave

ViPR

The Architect’s View

Related Links

Gartner ECB (External Controller-Based Disk) – 4Q2012

The Architect’s View

Related Links

Death of The PC (Or Mine At Least)

Google’s Hollow Cloud Storage Offering

The Architect’s View

Manipulating the Message – The Art of Marketing

Company N

Company E

The Architect’s View

Related Links

Data Storage and The Question of Trust

Trust

The Architect’s View