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.

Related Links

• Making Sense of Storage Vendor Growth
• Magic Quadrant for General Purpose Disk Arrays (Gartner)
• Q42012 Worldwide External Controller Based Storage Figures (Gartner)

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).

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.

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.

Related Links

• NetApp Exec Shakeup Names New Strategy, Technology Chiefs (CRN)
• NetApp Takes Build-It, Not Buy-It, Approach To FlashRay Flash Storage Strategy (CRN)

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).

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).

It’s easy to use the examples of server virtualisation and networking as reasons why thinking of storage virtualisation for SDS is only just around the corner.  However those assumptions are incorrectly made and as usual, storage is a special use case, that doesn’t quite fit the mould of compute and network.

Maintaining State

Whether we like it or not, storage is different. In a virtual server environment, the image of the server is held in memory, using a data image on disk as the means of maintaining state.  Only changes are committed to disk and these can be asynchronous in nature in order to improve performance.  If the physical server is rebooted and the in-memory copy is lost, it is simply reconstituted from the disk image and off we go again.  Moving virtual server around the physical infrastructure is simply managing data in flight.

In networking, data is transient across the network and doesn’t reside in the switch other than temporarily, as it moves between compute and compute, or compute and permanent storage.  The data is ethereal and the network was designed to be just that; capable of losing data during transmission, with high level protocols designed to manage that scenario.

Storage arrays (and storage in general) have to provide a different purpose.  It is the permanent record of data.  It has to be the part of the computing infrastructure that maintains state, even when the power is off.  That as we know presents special challenges.

Software Defined

For networking, “software defined” means splitting of the command plane and the data plane.  Simply put, the silicon in the network switch responds to management from an external device, directing packets as it is instructed to do so.  To carry this analogy into the storage world, we have to look at two pieces; the transmission of data and the storage of data.  For transmission, we expect to use Fibre Channel (FC), Fibre Channel over Ethernet (FCoE), iSCSI (IP SCSI) or perhaps something more bespoke like Infiniband or more proprietary like FICON.  The problem is, technologies such as Fibre Channel weren’t defined to expect a dynamic network.  A source and especially target device were expected to be static devices that joined and stayed in a network (or at worst, very rarely left).  A Fibre Channel network change was a disruptive event, RSCNs initially notified all devices in the fabric, which if occurring too often could cause fabric issues (over time vendors have minimised the effects of RSCNs).  For the networking component, SDS could mean a more flexible routing of data across the SAN.

However, the second thing we need to consider is the requirement to store data permanently, a concept that doesn’t exist in Ethernet networking.  It isn’t that simple to decide a that data volume or LUN now resides at the end of another connection and for traffic to be routed there.  What happens to the existing data?  How would data be moved and how would data integrity be maintained? Most important, what happens in a disaster scenario?  This is the hardest part of trying to work out what “software defined” means in a storage context.  Some vendors have used the idea of storage virtualisation or running as a virtual machine to represent this part of SDS.

Todays Examples

So does Software Defined Storage exist today?  In limited ways, I think it does.  One example is Hitachi’s Universal Volume Manager feature within the VSP platform, also known as external storage virtualisation.  This enables data to be written to an abstract device (which could be an internal disk or an external array) and for the control and data to be treated separately.  The array receives and writes data to the target device, but can be directed to write data to another device through the separate control plane.  This can even include (with Hitachi Availability Manager) redirecting I/O to a secondary device without requiring host interaction, but spoofing WWN addresses.  It can also mean redirection within the array using Tiered Storage Manager.  Incidentally, the image presented here shows a diagram I produced for Hitachi over three years ago.  It shows how the technology (which has changed names in some instances) can be placed into layers, in a similar fashion to the one Chuck uses in his presentation.  Good ideas never go out of fashion, don’t you think?

VPLEX is another platform I think meets what SDS could mean.  Data can be stored across multiple nodes, rather than statically in one place, with the ability to direct that control separately from the data path.

There are also other vendors offering products that fit some aspects of SDS.  Solidfire for instance, create an array of nodes that are managed using REST APIs.  The data and the control are separated from each other, with provisioning and management handled separately via API.  Other platforms like Nutanix do this too, although they are adding compute into the mix.

The Architect’s View

Software Defined Storage is a difficult term to pin down.  Today’s storage protocols, along with the need to ensure persistent storage of data whilst maintaining data integrity mean that the dynamic nature of SDS is hard to achieve.  There’s still a way to go before storage can be considered completely abstract enough to be termed “software defined”.

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).