Response: OpenStack Isn’t Our Savior from Lock-In or Support Costs — The Lone Sysadmin

Bob Plankers is making the point that the purchasing proprietary corporate software for virtualization (such as VMware vCloud or Citrix Cloudstack) has its own value by avoiding having to build & test your own software:

There is an attitude among some now that OpenStack is, or at least will be, our savior from vendor lock-in in the Infrastructure-as-a-Service cloud space, as well as something that will help corporations save a lot of money in licensing fees from VMware. While I see the potential I think there’s more to the picture.

This is true. HOWEVER…..

If you are building a cloud platform for 5000 to 1000 hosts, then considering VMware or Citrix is not a bad idea. It’s faster, easier and more reliable to pull some software off the shelf, read the manual and crank the handle and deploy a generic compute resource pool that you can fit your business into.

If you are building a cloud platform for 50K to 100K hosts then the numbers don’t stack up. The cost per host for VMware/Citrix quickly becomes unsustainable and using Open Source becomes perfectly viable.

Yes, a VMware implementation is more expensive in license fees and software support, but compared to hiring more people it’s a wash. You can spend $100K yearly on staff or you can spend $100K yearly on SnS from VMware (or Piston Cloud, for that matter, if you want commercial OpenStack support). For me, the tie always goes to the solution that involves fewer humans. Humans are expensive, take up physical space, require expensive, complex, and unproductive auxiliary support systems like payroll and vacation and bathrooms and laptops and air conditioning, are annoyingly subject to Metcalfe’s Law, and are often just as temperamental and hard to work with as VMware software.

When licensing software for more than 10K units, it’s more likely to be cheaper and better to use humans. Even better, you can design and more tightly specify your solutions to exactly match your business needs. Look at what Facebook has done with MySQL as a case study.

OpenStack Maturity

It might take another three to five years for OpenStack to achieve maturity (and that’s how long VMware needed to get stable and have the features) but OpenStack looks to be the most viable contender for cloud deployments at scale.

Given a choice between buying more humans or buying more vendor support I usually side with whatever solution has fewer humans, and automating the difference. And frankly, most organizations will opt to have a commercial support contract anyhow, thereby negating most of the SnS savings arguments.

Here is where I respectfully disagree. While it is complex and “hard” to manage a lot of human meat puppets to deliver projects, Amazon has proven that it’s possible. And Facebook. And Google. You business can also achieve these outcomes if you are smart and able to refocus on your IT resources.

Certainly, the MBAs of this world (now proving to be mostly worthless) would prefer to have zero workers and maximum profit. And Boeing has discovered that too much outsourcing causes it’s own problem (and buying commercial software IS outsourcing )

And that’s a major problem. It will take a few more years before you can “redesign” your IT Team to work in new ways. Until then, VMware is a good enough answer for more organisations to start on the “cloud journey” but I’m doubtful that VMware is part of the big future.

via OpenStack Isn’t Our Savior from Lock-In or Support Costs — The Lone Sysadmin.

  • OmarSultan


    One thought–the folks you name (FB, Google, etc) are essentially running their IT as a profit center (its central to how they make their money) which has a different investment model than the typical enterprise that runs their IT as a cost center (i.e. they make they money selling cars or little blue pills). I actually agree with the sentiment, but we need to be able to point towards successes in the enterprise to help it make sense to the MBAs out there.

    Omar (@omarsultan)

    • Etherealmind

      And I would maintain that we need smarter MBA’s who understand the company product and value. Today, all they know is ‘business’ and nothing about values and production.

      And we need smarter management who will engage with technology instead of claiming it’s too complicated for them. If managers can’t comprehend what they are managing they shouldn’t be doing that. It’s not acceptable to claim “it’s just business” – “it’s just technology too”.

      • OmarSultan

        Don’t disagree, there needs to be a meeting in the middle. Business execs should be conversant in IT as they are in other aspects of their business like the factory costs of call center efficiency metrics. However, IT folks also need to be better at showing relevance to the business–its a skill the senior execs for sales, manufacturing gain in order to reach the exec level, but IT execs still can struggle with. We in IT often couch the conversation in technical metrics–throughput, performance, etc–we need to get getter at linking back to the things exec care about–here is how this IT investment will help with customer sat or help us close deals faster. This is also where the consumerization of IT also does not help–where they can’t really see a whole lot of difference between consumer grade and enterprise grade and are left wondering why IT is so slow and expensive–I just bought an access point at Best Buy for $50 bucks and my nephew connected it up in 20 min–why does IT charge me $1000 and take to weeks. Now you and I know this is not an apples-to-apples comparison, but for someone focused on keeping customers happy and growing the business (which is frankly where it should be), it all looks the same and they are left with a mistaken impression. It is up to fix that or things are not gonna change–I know, I know, we should not have to, but that’s the reality in many companies.

        An analyst firm, Forrester, I believe, broke CIOs down into three categories: type #1 exist where iT is viewed purely as a cost and there is relentless pressure on the CIO cut expenses at all costs. #2 is where there is acknowledgement that IT is a key function and the CIO has a seat at the table for strategic planning, like the rollout of a new POS system, and then #3 is where the CIO has established a leadership position and can show the rest of the organization how IT can help lead the business. I am guessing most readers can figure out where their companies fit into that continuum. Progressing across the continuum is somewhat about the CIO themselves, but its also about the team behind the CIO providing the data to put IT’s contributions to the business in the proper perspective.


        • Tomas Fidler

          To the problem with Enterprise vs Custormer grade. That is indeed big problem…

          • OmarSultan


            There is a book titled “Real Business of IT” by Richard Hunter and George Westerman while lays out a good framework for communicating the business impact/value of IT investments–I gave it to my team as a summer reading assignment once. :)

  • Etherealmind

    Hey Bob

    Saw your response and appreciated your perspective. I agree about your differentiation on consumption versus production. The process of eating a hamburger is different from making a hamburger and should not be confused.

    At this point, VMware licensing is generously priced for VMware. It can be valuable for accessing proven software but requires significant capex to acquire. OpenStack lets you invest gradually over time using existing people and existing hardware – this is attractive to other businesses.

    I also think OpenStack will need some time to mature in the same way that VMware took many years to reach it’s current position. We shouldn’t measure today’s product too harshly because it may be too early to commit.


  • Massimo Re Ferre’

    “If you are building a cloud platform for 5000 to 1000 hosts, then considering VMware or Citrix is not a bad idea”

    “If you are building a cloud platform for 50K to 100K hosts then the numbers don’t stack up.”

    It is like saying that NASA will never buy an Audi A8 to go to the moon.

    The biggest telco in one of the European countries over here has just built a cloud with 4 (four) hosts. So many others with “dozens” of hosts.

    I don’t think Google, Amazon, Rackspace will ever use vCD or CloudStack… I believe it makes more sense to discuss what the remaining 99.9999999999% will be doing. But if you want we can keep talking about clouds with 400M hosts. Admittedly it sounds much cooler than a cloud with 12 hosts!

    Slightly off topic (but important to the overall discussion)… it will also be interesting to see how much of the market the 0.000000001% will be able to win and how much of the market the other 99.999999999% will get. It goes without saying it’s not going to be proportional. Of course if 0.0000000001% of the SPs will get 99% of the workloads (as some speculates) … this is a moot discussion.