What Is Converged Infrastructure? (a response to ZDNet)

Recently there was an article posted on ZDnet entitled “Converged Infrastructure vs Reference Architecture” which has collected a bit of attention and I wanted to add my thoughts regarding what I thought were some misleading conclusions.

If the article read as if it were written by an enthusiastic VCE employee, that’s because it was.  Now that’s not inherently a bad thing – the Vblock is a great product with much to be excited about (as my own blog posts reveal).  However, I walked away feeling as if the article did not accurately describe the market or the trade-offs involved with the various solutions.  I observed this article being shared multiple times but mostly from people with an agenda and not the more independent analysts.  In the spirit of full disclosure I work for a NetApp Partner where we offer FlexPod solutions.  Now let’s take a look at that article….


Let’s start at the beginning.  The first problem I have with the article is the title: “Converged Infrastructure vs. Reference Architecture”.  This suggests of course that reference architectures are NOT converged infrastructure and that customers must therefore choose between the two.  Is that truly the reality of this market?  Let’s start by defining what converged infrastructure is.  Here is a selection of excerpts from the Wikipedia definition of converged infrastructure:

Converged infrastructure packages multiple information technology (IT) components into a single, optimized computing solution. Components of a converged infrastructure solution include servers, data storage devices, networking equipment and software for IT infrastructure management, automation and orchestration.

Converged infrastructure is used by IT organizations to centralize the management of IT resources, consolidate systems, increase resource utilization rates, and lower costs. These objectives are enabled by the creation of pools of computers, storage and networking resources that can be shared by multiple applications and managed in a collective manner using policy driven processes.

If we look at the benefits of converged infrastructure we can further improve our definition:

Writing in CIO magazine, Forrester Research analyst Robert Whiteley noted that converged infrastructures, [combine] server, storage, and networks into a single framework, help to transform the economics [of] running the datacenter thus accelerating the transition to IP storage to help build infrastructures that are “cloud-ready”.[3]

Can a reference architecture do these things?  Could a FlexPod or VSPEX solution for example be used to create “pools of computing, storage and networking resources that can be shared by multiple applications and managed in a collective manner using policy driven process”?  That’s not only what they were designed to do but it is exactly how they are being used today.  In the case of the FlexPod – the market leading reference architecture – customers choose their own automation/orchestration solutions to best match their needs and business model ranging from Cloupia, CA, Cisco IA and more.  And of course these solutions automate and orchestrate across Cisco UCS  (compute), NetApp storage and Cisco networking solutions, leading to….tada!…..converged infrastructure (some real world examples in a bit).

Now a solution using the product model (such as Vblock) could indeed make the claim that their solution is “more converged” when it comes to procurement and support because you are dealing with one product and one vendor.  This is absolutely true, but it is quite different than suggesting that reference architecture is not a converged architecture, given the established definitions.  Furthermore, this overlooks the acclaimed support model NetApp has cultivated with their FlexPod partners which allows customers to make the initial call with the vendor of their choice, and then the call is seamlessly managed across other vendors in the stack as needed with shared ownership along the way.

It seems that someone is trying to bend the definition of converged infrastructure into a shape to better suit their own agenda as even solutions like HP CloudStack and IBM Pure Systems are labeled as somehow not being converged infrastructure.  In the spirit of this election year, I rate the suggestion that reference architectures are not converged infrastructure as “Pants on Fire”.  Let’s move on…..


The article opines that “…Reference Architectures are a great solution for a low budget SMB that is looking to introduce itself to the world of Cloud” and goes on to attempt to explain that only a “true converged infrastructure” is suitable for “serious contenders” looking for a “robust, scalable…private cloud infrastructure” and of course finally concludes that “the only such solution….is VCE’s Vblock”.  One wonders how this advertisement passed editorial review at ZDNet.

It is certainly true that the Vblock is well suited for high end enterprise/cloud environments, as well as that the FlexPod is very accessible for the SMB customer.  It is understandably tempting to limit these products to relative positions as a result, but the clear fact is that reference architectures like the FlexPod are being successfully used in world-class enterprise and cloud deployments.  I’ve only so much room here, but here’s just a few examples.

Accenture uses FlexPod to host a private cloud for SAP customers as well as a FlexPod powered cloud for the French government (G-cloud).  Euronet chose FlexPod for their central datacenter supporting over $50 billion in financial transactions in Europe annually.   Terremark (Verizon) is one of several hosting providers using the Flexpod, along with several other blue-chip names – one of which to be revealed this week.  In fact, if you look at the list of DaaS/IaaS and other service providers already using NetApp storage – it’s probably not a big stretch that several are FlexPod customers as well.  ING Direct and EGGER are among some of the larger international FlexPod customers, and if you’ve been a patient at one of the Cook County Health Care System facilities in Chicago, your patient records – along with another half million patients each year – are stored on a FlexPod powered datacenter.  So as you can see, just because the FlexPod may be more accessible to the SMB market does not mean that it can’t – and isn’t being used in a world class enterprise cloud offering.  Now ask yourself if EMC launched the VSPEX offering to focus only on the SMB market.  Here we see clear examples of a reference architecture providing world class enterprise cloud services well beyond the SMB space with more to come.  Pants still on fire.


On a recent trip to IKEA I found myself trying to match different parts of different sizes in order to scale out an existing (cabinet) solution.  It would have been nice if there were a document helping to show me which parts were designed to work together, in order to save me some engineering pain.  In this respect, I think that the reference architecture will replace the traditional “build-your-own” approach and gain further traction in the marketplace.  Now there is no denying that the Vblock – being fully “productized” – has several advantages.  There is much more engineering that goes into a Vblock for example than a reference architecture like a FlexPod.  And while there are clear advantages to this approach, there are also more parameters that one might have to operate within as a result.  For example, you might have to wait significantly longer to upgrade to the latest build/release until all the components have been fully engineered to support that release, as well as other potential boundaries and limitations required for support.  For many this will not be an issue, but for some it will.  There is less engineering that is pre-established for you in the reference architecture and yes this can increase the complexity of support, but it can also empower the customer with more flexibility – including being able to select the orchestration suite that best meets your needs and objectives rather than the one offering that is built into the product.  I’m told by a partner that one Vblock customer ended up dissatisfied with the solution and has placed four different FlexPod orders over the past 6 months.  There are a few trade-offs at play here and there is no “one-size-fits-all” solution.

The ZDNet article to me read as if there were only one viable solution and I just don’t see the market this way.  The Vblock is an outstanding product, but in my opinion so is the FlexPod – customers should take the time to understand these trade-offs and what works best for them.  I love being able to explain the value proposition of the FlexPod to customers and how they can choose among orchestration solutions that are the best fit for their goals.  Yes, there is more engineering put into a “full product” like the Vblock than references architectures – whether this is the right fit may depend on your goals, preferences and operational model.    At the end of the day what we all seek to do is to provide converged infrastructure in a manner which enables and empowers business to achieve new benefits from this utility model of computing, and there’s more than one solution that can effectively achieve this result for customers.

Now here’s my spin — I love what the FlexPod solution provides to us across hosted and on-premises IaaS/cloud solutions.  Cisco UCS + VMware + NetApp + choose your own orchestration + a converged support model = value for our customers and own own offerings.  It’s tremendously exciting to work with customers and market opportunities with the FlexPod and come up with solutions that provide phenomenal value and potential that has all parties incredibly excited.  At the end of the day, there are several different converged infrastructure solutions which are providing great value in the market place.  I don’t think that a reference architecture is somehow an incomplete or insufficiently “robust”  solution just because it is not fully “productized”.  In my opinion, a reference architecture — such as the FlexPod — can provide an outstanding basis for IaaS, private cloud, value and opportunity and I’m excited to have the opportunity to design solutions around it.


26 Responses to What Is Converged Infrastructure? (a response to ZDNet)

  1. jdooley_clt says:

    @BlueShiftBlog Wow, that might be the fastest kool-aid drinking I’ve ever seen, Kevin. 🙂

    • BlueShiftBlog says:

      @jdooley_clt Where’s the Kool-Aid? I admire the Vblock offering but to suggest nothing else is “robust” — that’s the real Kool-Aid IMO.

      • jdooley_clt says:

        @BlueShiftBlog Well, there really isn’t any other answer you COULD have, is there? Hammers find nails. It’s all good.

        • BlueShiftBlog says:

          @jdooley_clt was awkward to write as I admire several solutions, but ultimately didn’t care for how market was portrayed.

        • jdooley_clt says:

          @BlueShiftBlog Customers define the market, not the vendors. They will show you that the difference between CI and RA is real. You’ll see!

        • BlueShiftBlog says:

          @jdooley_clt I’ll agree that is is possible for a RA solution to NOT be fully converged, based on how it is engineered, but it CAN be.

        • jdooley_clt says:

          @BlueShiftBlog A customer, with work and investment, can make a CI out of an RA. Some don’t want to go to the effort and wish to buy a CI.

        • BlueShiftBlog says:

          @jdooley_clt Absolutely agree. It’s a trade off. Depends on skills/goals/ops etc.

        • BlueShiftBlog says:

          @jdooley_clt of course that burden doesn’t always have to be on the customer either. They can contract to build the framework.

        • jdooley_clt says:

          @BlueShiftBlog IBM’s been selling that line forever: “With enough services, anything is possible!” Let me know how your customers respond…

        • jdooley_clt says:

          @BlueShiftBlog And I don’t want customers to live with a trade off. Best components, best partners, fastest time-to-market, least risk.

        • BlueShiftBlog says:

          @jdooley_clt a valid contrast. I just don’t think it’s the ONLY path to enterprise-class IaaS/CI

        • jdooley_clt says:

          @BlueShiftBlog OK, I’ll give you that. Maybe we should make a list of things required for inclusion before you can be a “real” CI. 🙂

        • BlueShiftBlog says:

          @jdooley_clt doesn’t ITIL have that defined somewhere? 🙂

        • angryjesters says:

          @jdooley_clt @BlueShiftBlog sounds like a certain “product” I’ve been suffering with as of late…

  2. VCErichardb says:

    I don’t think its a question of whether its more or less robust – as we discussed previously, the difference is how you buy and consume it.   Engineering a different blend of RA for every scenario and application might be ‘interesting’ from a technology perspective, but if you’re a line of business or application manager – surely your focus has to be on reducing risk and complexity?

    • BlueShiftBlog says:

       @VCErichardb Thanks for responding Richard.
      Standardization can indeed be a means to reduce the risk and complexity.  This can be built into the stack, or a more custom solution can be engineered for your goals/environment.  Some clear trade offs here, but the bigger point I was trying to make is that there isn’t only one path to converged infrastructure providing world class IaaS.  Saying converged infrastructure is X and the only way to get there is also X, just didn’t sit right with me.

  3. BlueShiftBlog says:

    @ROIdude @gurusimran Thanks guys! 🙂

  4. BlueShiftBlog says:

    @mthiele10 thanks!

  5. Pingback: FlexPod In the Enterprise -- GE and More | Blue Shift

  6. Ed Grigson says:

     @BlueShiftBlog Nice post, and I agree completely with the tone of the original article. I’m not using either Flexpod or vBlock so don’t have any expertise (or agenda) but I’m bored of the mud slinging. The best way to ‘sell’  to me is to give me the facts and apply them to my situation, not bend the facts to fit. I have an interest in converged infrastructure but anything too one sided gets discarded in my mind as marketing, not information.

  7. HB says:

    Recently went the Vblock route having considered the other options specifically HP and Netapp. While on paper HP and Netapp were cheaper further investigation proved otherwise as our CFO scrutinised the long term OPEX. So why did I post a comment?
    Well I’ve had the pleasure of meeting Archie and had read his article as well as your response during our decision making process. No fact bending or sugared up marketing, VCE delivered exactly what they promised. Delivered the Vblock in 37 days and we were operational in under 48 hours. No way were we promised or even going to get that from a reference architecture!

  8. I for all time emailed this blog post page to all my contacts, since if like to read it afterward my contacts will too.

  9. I know this site gives quality depending posts and other data, is
    there any other web site which provides such things in

  10. Hi there, for all time i used to check web site posts here early in the break of day,
    for the reason that i enjoy to learn more and more.

  11. Without a doubt, after you buy Youtube views your task as the person who owns the route is not more than. A
    superior landing page on a website can helps to gain the trust of
    clients and keep them coming back again and again on your page.
    The best way to meet your potential customers’ needs is to
    give them a way to interact directly with you.
    com is concentrated on providing customers with exemplary purchaser organization collectively with a
    versatile technique to accessing organization info.
    The Content should be relevant and updated on a regular basis to generate traffic to your
    site. An app version is also downloadable which makes it easier for
    you to read, organize, and share the topics that interest you.

Leave a Reply

Your email address will not be published. Required fields are marked *