At the ripe old age of 5, will OpenStack simmer down?

7 Comments

I don’t normally go for these year-ahead predictions by tech big wigs (remember that before you hit send) but an item on this list by Rackspace CTO John Engates struck me.

Engates wrote that as OpenStack turns five, it’s getting boring — something that Derrick Harris and I, in our very non-scientific way, have been noting on the Structure Podcast for the last few months.

It’s important to note that he doesn’t see “boring” as a bad thing. Quite the opposite. He wrote:

“When a technology matures, it becomes less and less exciting. That’s where we see OpenStack going.”

He cited a Forrester Research brief coming out of the last OpenStack Summit in Paris that referred to:

 “a lack of excitement that comes with maturity. The Juno release [of OpenStack] addressed many challenges holding back enterprise adoption to this point and showed signs that 2015 may be the year its use shifts over from mostly test and development to mainstream production deployments.”

I’m not sure I buy this entirely. There are so many companies with big vested interests in OpenStack that there is bound to be contention as they try to differentiate [company]Rackspace[/company] OpenStack from [company]HP[/company] Helion OpenStack from [company]Red Hat[/company]OpenStack, and on and on (and on) down the line.

That is why the OpenStack Foundation’s DefCore effort is important. DefCore would ensure that different functions work the same way across OpenStack distributions so OpenStack doesn’t go the way of Unix, which ended up as a splintered operating standard with [company]IBM[/company] and HP and Digital Equipment Corp. and other companies all offering their own slightly incompatible versions of the operating system. That meant the promise of vendor independence was shot and customers got burned.

I’m inclined to agree with Battery Ventures’ Adrian Cockburn’s assertion that despite all these efforts, the technology will fragment even though the OpenStack brand will be widely adopted.

So, if DefCore succeeds, the dozen or so vendors with OpenStacks of their own will make up an ecosystem of interchangeable frameworks. If it doesn’t, some of the big players will go their own separate ways. Part of this may be wishful thinking on my part: There’s nothing like chronicling the growing pains of a potentially great technology or covering a good OpenStack spat every few months to focus the mind. And face it, the upside of people fight about something, is that they actually care about it.

Note: This story was updated at 12:34 p.m. PST to correct the name of the Foundation’s effort to enforce OpenStack interoperability. It is DefCore, not CoreDef.

7 Comments

Sean Roberts

Thanks for covering our efforts around DefCore. It is a critical OpenStack board committee. It’s implementation through the RefStack project is nothing short of amazing progress. Rob Hirschfeld, Joshua McKenty, and Troy Toman have put in selfless, long hours on making DefCore happen.

OpenStack will live and die by collaboration and competition in equal measures. The interoperability that the RefStack will provide OpenStack, will go a long way to insuring a healthy future.

Boring? I am excited about 2015. I think it’s the year commercial open source comes of age. See you there :)

Rag Ramanathan

I believe it should be Battery Venture’s Adrian Cockcroft (not Cockburn).

Kishore kumar Neelamegam

Too many hands spoil the broth. That is what is happening with Openstack and a need for a project like DefCore itself.

When so many super duper guys are involved in the board why can’t they have managed fragmentation efficiently.

5yrs and less adoption needs to be thought seriously.

The complexity of Openstack will only make Openstack a reference architecture/implementation and will drive enterprises to use the product engineering services firms to deploy Openstack if needed.

Which means the adoption will be slow as there is a jumpstart cost involved for busineses to get into Openstack.

I think the priority of Openstack is to make it less complex like Opennebula which makes enterprises adopt with ease than managing fragmentation.

The analogy isn’t fragmentation of Unix, but Openstack needs to thought as a linux kernel, with different distro’s taking shape as they wish. The core remains the same…

Lets wait and watch where it takes.

Ken Rugg

A couple of quick thoughts on this. First, I always find it troubling when people refer to development and test as if it doesn’t matter. Sure, applications have to get into production at some point, but investing in infrastructure to make your high priced development and QA staff more agile and productive can yield a very valuable return. Much of VMware’s early success was built on this value proposition.

Second, you really shouldn’t fear that DefCore will make covering OpenStack boring. There is still plenty of room for fighting, (and collaboration and innovation…) on the OpenStack projects that aren’t part of DefCore. Even if the project does create less differentiation in the core, interchangeable components, you’ll still have plenty to write about around the edges. In fact, if it drives more rapid adoption of the core, it might provide even more spats and growing pains for you to cover in projects layered on top of the core.

Keith Townsend

I agree that boring doesn’t equate to dying or non-movement. Network vendors are betting big on OpenStack as a mechanism to fend off VMware in the SDN space. For example, Cisco is making a huge bet on OpenStack. Boring is good. On the flip OpenStack could be the ever promising VDI which is much more exciting but never potential filling.

Comments are closed.