Rolling out the six-monthly "OpenStack, is it a thing?" post

OpenStack, perhaps the most hotly debated open source project of all time. So where does it stand today?

openstack photo

October saw the OpenStack Foundation hold its six-monthly summit, this time in Tokyo, Japan. I've been covering OpenStack since it was created, and in that time have seen it grow massively, albeit with a few stumbles along the way. (Disclosure: The OpenStack Foundation covered some of my travel and expenses to attend the event.)

The Tokyo Summit was an interesting event, interesting because, away from the U.S., it could be a far more internationally-centric event. Many of the case studies we heard about showed this, and it was refreshing to see some great successes from non-U.S. OpenStack users.

As can be expected, there were some product announcements at the show, but perhaps more important was the fact that the foundation seems to be focused on where the real challenges lie. A case in point is the announcement of certifications for OpenStack developers.

There is a massive skills shortage in the OpenSack world -- indeed it was somewhat obscene to see just how much talent-poaching was going on at the event. By developing a certification program, the foundation should hopefully deliver more skilled developers to fill the seemingly voracious desire for talent in the community.

Of course there were some product announcements, and a perennial concern for me is around feature-creep with the OpenStack project. The Tokyo summit saw a huge focus on software-defined networking. Talking with people at the show it seems that telcos are an increasingly important customer group for OpenStack and, given telcos real need for far more networking flexibility, it kind of makes sense that there should be a thematic push around this. That said, we have seen different thematic pushes over the years with OpenStack and some might suggest that it's more a case of the foundation clutching at whatever use cases look most likely to justify its existence or prove the success of the initiative than anything else.

The other issue is around vendors pushing their own commercial interests around OpenStack. At the summit, Rackspace announced Carina, a new container deployment offering within OpenStack. While this is something that Rackspace has created, there is a risk that they might try and push the community into commuting time and resource to develop a container offering. It's not that containers have no relevance to OpenStack -- indeed, a container offering may well be a smart place to move -- but there is a need to differentiate between what is an important move for one commercial entity within he OpenStack ecosystem, and what is good for the project generally.

In a panel discussion, I put it to some analysts from the APAC region that the biggest problem with OpenStack for their enterprise clients is in knowing what it actually is. Over the years, there have bee a few different messages -- initially it was all about building their own version of Amazon Web Services (AWS). As AWS has out-innovated all comers, that isn't a viable proposition and hence the message has changed over time. If there's one thing that enterprise CIOs don't like, it's a rapidly moving target when it comes to a technology platform.

HP's unfortunate shuttering of its own OpenStack-based Helion public cloud cast something of a shadow on the event. Unfortunately, many commentators were equating an HP disaster as being an indication that OpenStack is a disaster. That's an unfair conflation; HP's history in the cloud is a fraught one and that can't be blamed on OpenStack. I'd put that down to leadership and focus rather than platform choice.

There were some good customer statistics unveiled at the show. The number of deployments moving through to production has almost doubled since the end of 2013 -- up to 60 percent now. There is also a marked growth in larger organizations (those over 1,000 employees) using OpenStack.

Most interesting for me were some good customer case studies. One in particular stood out from Fujitsu. Fujitsu is Japan's largest IT services provider and has 100 data centers globally. The company is standardizing its IaaS service on OpenStack and is planning on rolling its service out globally over time. Europe is promised in the first half of 2016, with the U.S. in the second half. It was fortuitous timing for Fujitsu that recent EU decisions have created real concerns among non-U.S. customers about the risks around having their data with a U.S. vendor. Fujitsu, along with other non-U.S. vendors, has a strong answer for those customers.

Of course the old arguments still surface on an ongoing basis -- there was some discussion, yet again, about whether OpenStack should finally commit to building AWS' APIs. The thinking from some people is that, since AWS is the undisputed leader in the public cloud, and that organizations want a private cloud offering that works within the context of their public cloud footprint, that this is a critical need. That is a debate, however, that has been going for years and won't be resolved anytime soon.

Overall, OpenStack is clearly doing some really good things. It has customers, committed commercial partners and a huge number of developers that live and breathe the project. The foundation and the board need to try and chart a path which sees them control some of the over-exuberance of some vendors, but that is always a very difficult task to handle. While some very outspoken commentators call OpenStack a "dead duck," it is obvious after spending a few days in Tokyo that OpenStack is well and truly alive and kicking.

As someone said to me at the event, Linux, perhaps the most successful open source project in history, took well over a decade to really get to scale. OpenStack is only half as old as that so I'm more than prepared to give it the benefit of the doubt at this stage.

This article is published as part of the IDG Contributor Network. Want to Join?

The march toward exascale computers
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies