Thoughts on OpenStack ‘hype’ vs ‘reality’ post

@cloudpundit’s (Lydia Leong‘s) recent Gartner article attempts to dispel hypes around OpenStack. Naturally, it wasn’t received by many – both from and outside of OpenStack community.

As a OpenStack enthusiast, it would have been easier to dismiss that as a FUD. However,  I think it is in the best interests of OpenStack project to consider actionable items. Thanks @cloudpundit (and your recommendations at the end were right on)! Here they are:

1) Publish data on size and stability of large scale deployments:

Concerns about maturity of the OpenStack can only be addressed by sufficient customer feedback. There are at least a handful of ‘large scale’, ‘enterprise’ deployments of OpenStack. Deployment statistics and user feedback from such large deployments would re-affirm OpenStack’s maturity.

2) Publish road map of product features:

Having a road map for the next 18 – 24 months ahead gives good confidence to the customers to take bets on. Moving to OpenStack Foundation is expected to give such clarity; this moves feature planning from project (nova/ swift) level to product (OpenStack) level.

There are a few that appear to unsubstantiated. Lydia, questions to you:

1) What exactly is the ‘hype’?

I couldn’t understand what exactly is the hype is and what is it about – is it the popularity of OpenStack or is it the buzz around OpenStack Foundation? To me, there is no more buzz than any reasonably successful open source product (and the geekiness around it). I also don’t think any CIO will go for OpenStack just by so called ‘hype’.

2) Do you have any data from current large scale users of OpenStack on lack of maturity?

I would be interested in knowing the data points. Obtaining deployment numbers from partners was very difficult – of course they had their business reasons not to reveal. However, if you have come across such data, it would be very helpful.

3) When would you consider a CMP to be matured?

Is it by customer adoption numbers? Sizes of deployments? Feature  set? I don’t think one can compare check against AWS features for maturity. While OpenStack feature set can be expanded, I would rate all of the pre-Essex features (nova, swift, keystone) to be mature.

4) Technical Evaluations:

How is evaluating a CMP uniquely applicable to OpenStack? No CIO would consider a CMP without an evaluation; if so why only for OpenStack?

5) Open vs Interoperable:

Where does OpenStack documentation/ marketing claim Interoperability? In fact, one of my hopes  is that OpenStack standardizes cloud APIs, rather than aligning with a mess called AWS APIs. Your article seem to consider open and interoperable synonymous.

6) Open vs Vendor:

I do not see the OpenStack community as a cliques of vendor-employees. Particularly after OpenStack Foundation is live, it is an open source project with a lot of participating vendors and an eco-system of vendor offerings around it.


One comment on “Thoughts on OpenStack ‘hype’ vs ‘reality’ post

  1. I’ll provide an answer in brief, so don’t take this as a comprehensive reply.

    1. Hype has a life of its own. Amazon, for instance, did not create the hype around cloud computing, and yet it existed. Gartner publishes something called a “hype cycle” which is basically our barometer of how the typical man-on-the-street IT guy thinks about a technology. OpenStack has more hype, and more disconnect between hype and reality, than any OSS I’ve seen in a dozen years as an analyst. No one says these kinds of things about, say, Hadoop. While Linux got its share of hype, by the time it caught the popular imagination, it was well on its way towards mainstream.

    2. We talked to numerous people who either recently tried OpenStack or were currently using it (along with many vendors involved in OpenStack) and we watch the mailing lists as well. We don’t cite names in the report because nearly all of those conversations were under the guise of client confidentiality. Downloads are high; POCs are not; significant deployments are limited.

    3. Gartner has a formal framework for this, encompassing both business-type factors (strength of vendor support for an OSS distro, for instance), and technical-type factors (likelihood of bugs, for instance). A practical way to think about it might be, “Can Joe User deploy this and maintain it without frequently calling on his vendor for support / going to the community for help, running into bugs, performing maintenance that is high-risk (such as a fragile upgrade procedure), or having his tools break from version to version?”

    4. You’re wrong. One of the reasons that I published this note was that I was getting people calling me explaining that Management had decided they were going to adopt OpenStack without an evaluation, thanks to something somebody had read.

    5. For interoperability, you need interop between versions, between distros, and between other IT systems. For openness, Gartner has a formal framework for evaluation.

    6. I agree that OpenStack is not a set of cliques of vendor employees. I do think that vendor priorities drive significant decisions, though, especially with the “come with code” philosophy.

    I think, by the way, Rackspace’s blog post to the community upon the Foundation handover is right on the money. It implicitly acknowledges the issues that I raised in my note, and provided sound wishes for the future.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s