I’m something of a political junkie. Independent of the outcome, the results of the presidential election were a letdown in that the game was over. But notwithstanding the fact that I still read Nate Silver religiously, I rarely resist the impulse to try and quantify some social phenomenon. And there are more elections coming soon – specifically, the OpenStack foundation.
So this seemed like a reasonable time to take a snapshot of the membership of the OpenStack Foundation. Who are these people and what are their affiliations? What if anything does it say about the use and uptake of OpenStack cloud? What doesn’t it say? And, like any data, what other useful questions does it raise?
The membership data is here; you can see my listing here. As you’ll see, there’s only a little information. And if you’re not listed, or don’t like your listing, please go here and start messing up my data.
Now, let’s be clear: this isn’t science, and it’s not that precise a scorekeeping exercise, either . My methodology was mainly to normalize names, so that Hewlett Packard, HP, and HP cloud services are all counted as HP, and then assign most of those I recognized to some sectors I created. Don’t like it? Here’s the source data; please feel free tell me what you would do differently.
Now, my observations.
OpenStack: Democracy or Republic?
I don’t actually have an interest in settling that one, given it’s been messy since the Virginia Compromise, and filibusters are still with us. But the easiest data to scrape from the membership registry is company names, so here’s what we saw (all data here refer to individual membership)
As has been said often been noted, OpenStack is dominated by 3 companies: between them, RackSpace, HP and Dell have roughly 3234 of the 6710 or so individual members, or around 48%. My first observation, and the first call to action, is the number of members who listed no company at all – they constitute 20% of the membership.
Here's the question: Is open source participation still the equivalent of samizdat? Are these people independent swing voters, or are they protecting their identity, and if so, from what threat? Outside of the West, this is a serious question; of the 68 people who wrote their names in Chinese characters, 42 had no affiliation listed, a meaningful difference from the overall populace.
When you look at the named company distribution, excluding those people who listed no affiliation, the lopsided gerrymandering of big companies is even bigger – 60% of members naming organizational affiliations are from the three big players. This makes the unaffiliated vote somewhat more ironic. Does anyone think for a moment that each of the thousand-some individual members from HP and Dell is OpenStacking actively? For companies relatively underrepresented, such as IBM or EMC, they’re ceding the ground unnecessarily. Red Hat couldn’t send out the same email that Rackspace did to get people onboard? The anonymous affiliates are messing up the denominator – an invitation to filibuster, I’d say. Better to get more of your colleagues to join and affiliate openly, even if they’re only there to vote.
Another disclosure: Mirantis ranks 4th in the stack rank of affiliations (5th if you count 'no affiliation'). While this impressed my Mom (I asked her, she said so), our weight in the OpenStack ecosystem is comparable really only to Nebula, the only company about our size/shape in the ‘Next 10’, which I counted including roughly 600 members. Bragging rights: we’re bigger than Cisco, IBM, and EMC (maybe for no longer than it takes them to read this post and send out an email).
|4. Mirantis||104||9. Nebula||50|
|5. Cisco||85||10. Canonical||43|
|6. IBM||76||11. VMWare||36|
|7. EMC||69||12. Red Hat||35|
|8. SUSE||59||13. Dreamhost||33|
Who counts in the OpenStack community?
Aside from representational distortion of companies with well-organized employee teams, another thing I saw in this data is that the population of companies with exactly one member dwarfs the ‘senatorial’ model where each company counts once, independent of population. Of the roughly 990 company names, 825 or so had exactly one member. So, with 1000 or so companies each having at least one person paying attention to OpenStack, it seems like a good start to attracting broader interest in the project. We’ll look at what sectors they come from in a moment, but I want to say a word in defense of these individuals who may or may not have the rest of their company as ‘on board’ with OpenStack, as, say, Rackspace.
Recently, Thomas Goirand of GPLHost commented on a post in the OpenStack Linked In group (membership required) on his analysis of the project Git logs. He found 171 unique domain names and 550 contributor email addresses – and then observed that the “real numbers” are good so there’s no need to talk about thousands of developers and hundreds of companies. I demur, of course. More to the point, I used to hear guys in one of the Apache projects refer to user-list readers who did not subscribe as ‘lurkers.’ Snark notwithstanding, that’s not quite fair. You don’t have to commit code or subscribe to a list to make a project successful – those are necessary, but not sufficient.
What kinds of companies are into OpenStack?
Let’s set aside the pie-chart-denominator tomfoolery for a bit to reflect on what we might observe in the data with reference to OpenStack adoption in various segments. Here, more so than the affiliate census, the company segmentation is somewhat subjective. I don’t pretend that the categories are either mutually exclusive or exhaustive, though I have tried to make them accurate. I classified as many named affiliations as I was able. I came up with about 420 different companies I could identify and assign to a sector, representing some 4700 members of the 6710 or so publicly listed. This amounted to all but 8 people (2 each) at companies with 2 members, and a more significant 570-some at 1-member affiliations whose names I didn’t recognize.
So what does it mean to the adoption of OpenStack that there are 44 members whom I assigned to the Semiconductor sector? One claim that one often hears about deployments of OpenStack cloud is that ‘it’s not enterprise enough’. This is a pretty boring complaint, particularly when I hear it from writers in the mainstream media. What I think they mean is that companies who are not serious about making technology a differentiator won’t read their magazine. Are the semiconductor companies using OpenStack using it in ways that can’t be replicated by, say, Retail? Hello, Amazon?
Another critique of OpenStack uptake that I’ve heard is that that it’s not ‘mainstream IT’, as if it’s a bunch of crazy open source college interns who are there for the free sodas. We’ve talked to some big semiconductor companies about OpenStack, and the people we talk to are in IT. I think this is likely the case at EMC, Inventec, Sony, Orange – or any of the other companies with individual members. One thing that companies squarely in the tech sectors have learned that they ignore dog-fooding at their peril. So you can be assured that companies we talk to who are investing in OpenStack as a target market (they’re on the Foundation Board of Directors) are using it to help fly their own airplane.
Finally, you might object that ‘It’s just some guy in a cubicle, a hobbyist. If you look at my classifications, you’ll see I listed media companies like Pearson as SaaS businesses, rather than manufacturing, or publishing. Perhaps I’m getting ahead of Pearson (or so I hear from friends who work there), but SaaS is where they will live or die. The point is this: listing EMC as a Storage Hardware vendor is in the end an arbitrary distinction with respect to their mainstream commitment to OpenStack.
Individual interest in OpenStack matters
OpenStack Foundation elections are coming. Nominations are closed, and there are some solid candidates. What’s cool is not merely that these are some sharp guys (though it’s not cool that they’re all guys but one. Really). It's that they come from a range of sectors, big companies and small, and around the world. A special shout-out to Trung The Nguyen of DTT in Vietnam: "Trung and DTT are particularly interested in OpenStack for the strategic breakthrough it can bring to Vietnam – a developing country that must rely heavily on IT to find its way to growth."
So have a look at these folks, sign up for the foundation, and get the people at companies you know who have a stake in the future of OpenStack and cloud to do so as well.
Don’t like it? Here’s the source data, and please feel free tell me what you would do differently.