Home | Opinion | Why EMC Should Have Bought IBM’s Server Business
Why EMC Should Have Bought IBM’s Server Business

Why EMC Should Have Bought IBM’s Server Business

40 Flares Twitter 24 Facebook 2 Google+ 3 StumbleUpon 0 Buffer 9 LinkedIn 2 40 Flares ×

By now the news will be out – Lenovo are buying IBM’s X Server business.  But what would happen if things had been different and EMC made the first move?

I’m sure many of you are thinking that this strategy would make no sense, but bear with me for a moment.  Firstly let’s look at EMC’s business as it stands today.  The storage business is being attacked on all fronts and it could be argued that EMC is at an inflexion point.  We see converged and hyper-converged solutions increasing in popularity.  The former means EMC is in competition with Hitachi and HP, both of whom sell converged systems in which they own the hardware.  For the latter, these solutions don’t even require separate storage hardware.  The move to convergence means the mega-array (e.g. 4PB VMAX solutions) are less attractive.  The economies of scale in deploying large monolithic infrastructure is being outweighed by the disadvantages including difficult change control process, managing downtime, balancing performance and migrations onto new hardware.  The storage pendulum is swinging firmly back towards distributed storage again.

How could EMC counteract this?  They have started already, as we know.  From a purely storage perspective, they are attacking the all-flash market with XtremIO, admittedly from the position of follower rather than leader.  We’re starting to see the emergence of software-only solutions too.  These include ScaleIO, providing distributed storage, and talk of virtual machine versions of existing products.  There is also a move to sort out the management/orchestration piece with ViPR.  Note that some of these pieces have already been delivered by other vendors, notably HP with StoreVirtual and their OneView software.

What would server hardware bring?  EMC already knows a lot about the hardware business, selling storage arrays and being part of the VCE coalition.  However that partnership is somewhat rocky and Cisco could easily go it alone, having acquired Whiptail to add the storage component to some of their offerings.  This could start to chip away at the high-end sales of VCE and allow Cisco to develop high-IOPS solutions without EMC.  If EMC acquired IBM’s storage business they would get the ability to deliver their own converged solutions.  They could even buy Brocade to fill out the fabric piece and have no dependencies on other partners at all.

The Architect’s View

EMC are in both a very strong and weak position.  They own most of the jigsaw pieces and could have owned a significant part of the converged solutions business.  With a bit of boldness, they could have acquired server technology and transformed themselves into a provider of private cloud solutions, delivering the hardware and potentially in the future, software versions of their existing products.  ViPR could sit over the top, managing the whole thing, giving EMC what they want more than anything – account control.  Future acquisitions then simply slot in and offer services where there are gaps, such as scale out object storage.  Alternatively, they could do nothing and assume the storage business alone will sustain them.  However, many companies that have come and gone before assumed their business model was impenetrable and suffered as a consequence.  We only have to look to Nokia and BlackBerry to see how quickly the fortunes of the market incumbents can change.

 (Note: this post was written before the Lenovo announcement – literally minutes before – and has been amended to reflect current news)

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).

Copyright (c) 2009-2014 – Chris M Evans, first published on http://blog.architecting.it, do not reproduce without permission.

About Chris M Evans

Chris M Evans has worked in the technology industry since 1987, starting as a systems programmer on the IBM mainframe platform, while retaining an interest in storage. After working abroad, he co-founded an Internet-based music distribution company during the .com era, returning to consultancy in the new millennium. In 2009 Chris co-founded Langton Blue Ltd (www.langtonblue.com), a boutique consultancy firm focused on delivering business benefit through efficient technology deployments. Chris writes a popular blog at http://blog.architecting.it, attends many conferences and invitation-only events and can be found providing regular industry contributions through Twitter (@chrismevans) and other social media outlets.
  • http://ewan.im/ Ewan Leith

    It would certainly of been an interesting move, but I think it would have been a backwards one.

    Most storage companies are already using an OEM like Xyratek or direct contract manufacturing, but I doubt any of them have a desire to own the manufacturing process itself, especially for very low volume (compared to servers) storage hardware.

    I think it’s more likely we’ll see the EMC Isilion (or VMware VSAN) type VM deployments becoming more common, running on any certified hardware regardless of manufacturer, and that software layer is the future of a typical EMC storage deployment.

    • http://thestoragearchitect.com/ Chris M Evans

      Ewan

      How do you reconcile that view against how successful Cisco has been with UCS?

      One thing I didn’t mention in my haste was the whole VMware benefit. In my experience, EMC has been a company that wants to own everything, so putting their technology (virtual or not) onto or in with other vendor servers etc would be against their character. That means a really big personality change for the company. Would that be harder to achieve than spending a mere $2.3bn on their own server platform?

      Can’t say I know for definite but it’s an intriguing set of possibilities.

      Chris

      • http://etherealmind.com Etherealmind

        IBM server revenue in Q3 2013 was $2.8B according to IDC while Cisco was only $599MM and the market shrank overall by 3.7%. Specifically, mid-range & high end server sales shrank by around 20%.

        Being successful in a shrinking high end market probably isn’t something that Cisco or EMC want to be a part of. Their business models are tailored to high margin products that need expensive sales people, long supply chains and expensive mtce contracts.

        Sales of ODM servers grew by 45% suggesting that commodity products are progressively changing the server market.

        • http://thestoragearchitect.com/ Chris M Evans

          Greg, fair points. However for EMC, their business is built on high margin, using commodity devices – but their business is likely to be eaten away by other solutions. When EMC sold (for example) VMAX and VNX as a package, they could make good margin on that physical platform; I don’t think people will pay as much for software-only solutions, when other solutions are coming along.

          So what is EMC to do? I’d be interested in your thoughts.

      • http://ewan.im/ Ewan Leith

        UCS is a great product, but has Cisco been that successful with it?

        It’s annual revenue from UCS appears to be $2 billion http://www.bradreese.com/blog/12-9-2013.htm and has roughly 5% market share. That $2 billion is just 4% of Cisco’s overall revenue.

        The main benefit appears to have been to create a captive market for their Nexus devices, which remains a very high margin, very profitable product range.

        In comparison, Quanta are now manufacturing something like 15% of servers – if EMC really wanted to sell branded servers, it would be a much simpler task to contract with them than to buy a server business http://www.theregister.co.uk/Print/2013/03/12/quanta_cloud_hyperscale_server_business/

  • Alistair Stewart

    Using vanilla motherboards from Intel or Supermicro (e.g. EMC Isilon nodes) is quite good enough for most of EMC’s needs. Why try to compete in the low-margin server business when, in my experience, most customers seem to be happy filling their data centres with HP blade servers?

    • http://thestoragearchitect.com/ Chris M Evans

      Alistair, what happens if converged infrastructure takes off in a big way and end users want to “modularise” (ugh, apologies for the Americanism)? Does that mean EMC have a shrinking business, especially if VCE disintegrates? I’m speculating but trying to see where things might go in 5-10 years time.

      • Alistair Stewart

        You mean building the storage into the blade chassis and having compute and storage resources together? In that case you’re going to have a number of islands of storage, one in each modular sever which is not good unless you’re great at capacity planning or can share the capacity between servers. In that case, why not just separate the storage from the servers and share it between all of them.
        Most of the growth in the storage market is in unstructured data which doesn’t generally require the low-latency of a Fibre-Channel SAN so Isilon was a great acquisition for EMC a couple of years ago. There is still a market for distributed (and monolithic) structured data (databases) which can work well in a modular architecture, but this is not the same sort of growing market.

  • alpharob

    Disagree. The X86 server business has followed the path of other tech and reached true commodity (or well on the way). Recall there were many drive manufacturers 25 years ago, many PC manufacturers. There’s probably a whole bunch of business logic in dumping X86 servers, focus on higher margin, higher growth segments. Something like that. But IBM was very smart in dumping PCs when they did.

    • http://thestoragearchitect.com/ Chris M Evans

      Alpharob

      IBM changed to become a services business rather than focus on hardware (although have retained some of their hardware business, that at the time may have seemed strange, like the mainframe). Where do you see EMC going then? They can’t convert into being a services business and storage is going commodity too, just like servers. So, in 5-10 years time, EMC will have the same commodity business as IBM.

      Thoughts?

      • alpharob

        I’ll address some other points later, but regarding “changed to become a services business rather than focus on hardware” , read the smackdown in the comments section here: http://forums.theregister.co.uk/forum/1/2014/01/22/should_ibm_sell_its_storage_business/. Hardware will drag along quite a bit of services and software with it.

      • alpharob

        Well, where is HP and Dell going for that matter? EMC has one of the strongest M&A groups in the industry, perhaps more cloud acquisitions?
        British researchers just announced Flexigrid greatly increasing throughput using existing equipment. I suppose it is only a matter of time before bandwidth gets a lot cheaper. When that happens it will make a lot more sense for small and mid-size biz to move to the cloud and further eroding IT services and purchases for in-house. More and more I’m working with people with un-pronounceable names. From a business perspective it makes all the sense in the world to spend less in mindless IT tasks, outsource that to cheap labor pools. I’ve told all my kids to focus on something that can’t be outsourced and we have chats about what that would be.

        • http://thestoragearchitect.com/ Chris M Evans

          Dell I couldn’t say, but HP are focused on moving towards tightly integrated solutions; not just packaging compute/network/storage. Similarly Hitachi. If converged is really to take off, where does that leave those without a converged portfolio? They stay firmly at the packaging camp.

          Today EMC’s biggest competitors are NetApp, Hitachi, HP and IBM. Only NetApp are a pure play storage company, relying on reference architectures and no solid flash strategy. They have problems of their own. The others are selling storage as part of solutions. EMC II, the business excluding Pivotal and VMware, remains storage only with their VCE coalition for converged solutions and components they don’t own.

          This post was the basis for a discussion on where EMC heads in the future in terms of how the storage business will evolve away from expensive hardware solutions. I honestly don’t think EMC have a chance in hell of selling software only solutions at the same margin they made on hardware. There are too many competitive and open source solutions out there for EMC to acquire them all and software is the easiest thing to bring to market. There is also massive risk from AWS, Azure and the associated cloud storage gateways.

          It will be interesting to look back in a couple of years and see how things developed.

  • Morales

    It would be to much crap to take over for EMC? They aren’t really in the “cheap stuff business” so they would have to cut large parts of the product line. They could just OEM a server of EMC class and ride the investments of cheaper server manufacturers to drive their less demanding software.

  • Jesse G

    Been an EMC employee, contractor, and I’ve also worked for non-emc both storage and server vendors.

    As a rule, i think EMC has already gotten too far away from what made them great in the first place, pure storage. Their insistence on dabbling in software, appliances, and by moving to more commodity architectures, has made them weak, IMHO.

    I would love to see EMC get back to being the best at pure Storage with all the bells and whistles, and stop with this constant move to be mediocre at everything they can touch.

  • Pingback: Why EMC Should Have Bought IBM’s Server Business()

  • Pingback: EMC and VMware Working on Project Mystic for Hyper-Converged Infrastructure | Architecting IT Blog()

  • Rich Cramer

    I have a very different perspective on this while agreeing with Chris Evans; some part of the storage array business is going to be taken over by software-defined-storage (using software like Maxta) which run on physical servers so EMC would have been smart to buy IBM’s x86 server business. However, they can still access it through their partnership with Lenovo.

40 Flares Twitter 24 Facebook 2 Google+ 3 StumbleUpon 0 Buffer 9 LinkedIn 2 40 Flares ×