The bi-annual OpenStack Summit got kicked off this morning in Austin, Texas. I’ve attended every Summit for the past three years and this one is bigger than ever with over 7,500 attendees from some 61 different countries based on pre-registration data. Yes, this means that roughly 1/3rd of the world’s countries are represented at the Summit, assuming all those people show up.

Attendees are coming from about 1200 different companies, which means that (a) on average companies are sending good-sized teams to the Summit, and (b) all those attendees aren’t just from Summit Sponsors, since the number of those is just* 116. (*Though quite large in terms of the number of sponsors for a conference.)

Attendee roles are fairly diverse as well, with attendees self-identifying as the following:

17% – Cloud Architect
12% – OpenStack User
11% – Operations/System Admin
11% – Upstream Development
10% – CEO/CIO/CTO/IT Manager
9% – Product Strategy
9% – Business Development/Marketing
8% – Cloud Application Developer
7% – Product Management
1% – Media
6% – Other

Keynotes Day 1

As per usual, there are two keynote sessions at the event, with the first day’s keynote (roughly) focused on the OpenStack community and the second day’s keynote (again, roughly) focused on the OpenStack project itself.

Three key themes jumped out at me during this year’s day one keynote:

1. Enterprise.

The “Enterprise” was a central theme at the keynote this morning, perhaps best illustrated by the fact that both Gartner and SAP were, for the first time, keynote presenters.

This rubbed some folks the wrong way, and there was quite a bit of meta conversation on Twitter.

In general though, the answer to the question I posed in my tweet above (aspiration or reflective) is “both.” These keynotes tend to be both lagging (0-6 months) and leading (12-24 months) indicators of where OpenStack is going. Put another way: OpenStack is maturing and becoming of interest to increasing numbers of enterprises. At the same time, the Foundation sees this in the annual user survey data and begins to direct resources and the community to go after enterprise opportunities and the trend solidifies.

A good example of this is NFV. When we first heard about this at the Atlanta Summit it seemed to come from nowhere. Now it’s the driving use-case for OpenStack at the moment. In two years’ time, we’ll be seeing a lot more traditional enterprises gracing the keynote stage, especially since:

This is data from the user survey. Someone on twitter then made the leap to say “and 65% of those [Fortune 100] are in production, quoting another stat from the user survey. Unfortunately statistics doesn’t work like that, but I’ll see if I can get this number from the Foundation.

2. Telecom.

I mentioned this above, but NFV (network functions virtualization) has emerged as a killer app for OpenStack in the telecom industry. Kudus to the Foundation staff and community for seeing this opportunity and organizing around it.

Telecom and NFV played a dominant role in the keynotes. AT&T; delivered the first case study with a presentation on their journey to cloud. They went on to win the OpenStack SuperUser user-of-the-half-year award, following on the heels of another telco from last time–Japan’s NTT.

A customer logo slide presented during Red Hat’s video presentation was also pretty telco heavy.

During the media conference after the keynotes, OSF executive director Jonathan Bryce mentioned a great deal of functionality that has been built into OpenStack to support NFV. How the community will reconcile the needs of two large and demanding communities (ie enterprise and telco) remains an open question and will be a challenge.

3. Heterogeneity a.k.a. Diversity

A final theme this morning was the heterogeneity of IT environments and how that plays to OpenStack’s strengths. I see this as an evolution of the OpenStack-as-API or OpenStack-as-Glue marketing messages we’ve heard in previous keynotes.

It is certainly the case that both telecom and enterprise IT organizations are extremely heterogeneous. But public cloud environments and their progenitors –web-scale environments run by companies like Google and Facebook–are much less so. In adopting this as a key marketing message, OpenStack is telling users you can have your cake (complexity) and eat it too (cloud). Certainly this message was echoed by some of the keynote presenters. AT&T;’s Sorabh Saxena for example, credited their success to OpenStacks ability to help them manage the extreme heterogeneity of their cloud offerings. But I think that this can probably be taken too far, and that a key differentiator between successful and failed private cloud initiatives will be the degree to which heterogeneity is created.

Speaking of which, Mirantis’ Boris Renski gave a nice (and bearful) presentation on the difference between successful and failed cloud initiatives. In it he pointed out that people and process moreso than technology defines success and failure with cloud. His example focused on the operations team and how traditionally skilled ops teams (ie VMware sysadmins) won’t run a great cloud. Going back to my previous point, I think this starts earlier with architects as well. A cloud architect who knows how to streamline an environment for elasticity, scale and efficiency (in part through simplicity and homogeneity) is not necessarily the same as the enterprise architect who has reigned over a one- of-everything regime of J2EE hell.

So, while I fully support diversity as a community goal, I think it needs to be taken with a grain of salt in the cloud.

More Tweets

Here are more interesting tidbits from this the day 1 keynote.