OpenStack has a fantastic community, which may be its biggest problem. Matt Asay explains.
Cloud may feel like old news, but most enterprises have yet to do more than scratch the surface on cloud adoption, according to a new Goldman Sachs report. In 2014, cloud accounted for 5% of the $300 billion enterprise IT market, a number that will climb to 11% by 2018.
This is good, and a testament to Amazon Web Services (AWS) leadership. But clearly, more needs to be done, which is why it's disappointing that OpenStack doesn't appear to be doing much to take the cloud forward. If anything, by enabling overly complex private clouds, OpenStack has taken the industry backward, not forward.
What sells in the cloud
The cloud is all about convenience. Early on, many thought cloud would win because it was cheaper. This hasn't been the case. Not only is cloud not necessarily cheaper, it's not even more predictable in terms of costs.
What it is, is convenient, giving developers a flexible way to build applications that scale without having to involve corporate bureaucracies along the way. Such convenience has led to cloud taking a sizable chunk out of enterprise IT spend, as a new report from Goldman Sachs finds.
Sizable, yes, but still relatively small.
While cloud spending growth vastly outpaces the general IT market — cloud sports a 30% compound annual growth rate (CAGR) through 2018, compared to traditional IT's 5% CAGR — it will still barely top 11% of all IT spending by 2018.
This suggests that though the report calls out EMC, IBM, and Oracle as vulnerable to the cloud computing tide, they still have time to regroup and rethink their responses to cloud. Oracle, for its part, spent years laughing off the cloud. Its miserable earnings reports, quarter after quarter, finally prompted the enterprise IT giant to get serious, but the transition has been painful.
The answer for several of these behemoths, at least in part, has been OpenStack. But OpenStack may not be the cloud enterprises are truly looking for.
More companies giving OpenStack a try
The community growth around OpenStack has been stunning, leading to enterprise interesting in running it in production.
The primary reason for this adoption seems to be a quest for greater innovation. As the November 2014 OpenStack survey data indicates, the top reason for embracing OpenStack is a belief that it will increase "the ability to innovate."
If only that were as true in theory as in practice.
The reality of OpenStack is that it remains exceedingly complex, largely due to the very thing that makes it so appealing:
Theory vs. reality in OpenStack land
No one knows OpenStack better than Cloudscaling's Randy Bias. As one of early OpenStack pioneers (and founding board member), Bias understands all the potential of OpenStack, as well as its limitations.
So, when he declares that "OpenStack is at risk of collapsing under its own weight," OpenStack adopters need to listen.
In a recent blog post, Bias calls out significant tension between "The original vision [of] the OpenStack community delivering a tightly integrated release focused on basic infrastructure services on a 6-month release cycle" and "the inherent inclusivity of the OpenStack community [coupled with] people's wildly differing interpretations of the word 'cloud'." In other words, while OpenStack seeks to harmonize a mess of competing interests in a truly open community, what it desperately needs most is "real product management and product strategy leadership."
Or, as I wrote nearly a year ago, OpenStack needs a "a dominant vendor like Red Hat to ensure that the engineering, marketing, and sales of OpenStack get handled in a professional manner."
Without such leadership, would-be OpenStack adopters like David Laube, head of infrastructure at Packet, learn the hard way that community doesn't necessarily deliver the convenience enterprises expect from the cloud:
"Over the course of a month, what became obvious was that a huge amount of the documentation I was consuming was either outdated or fully inaccurate. This forced me to sift through an ever greater library of documents, wiki articles, irc logs and commit messages to find the 'source of truth'. After the basics, I needed significant python debug time just to prove various conflicting assertions of feature capability, e.g. 'should X work?'. It was slow going."
Laube was willing to dig into the code and mailing lists to deeply understand OpenStack. Most, however, will not. But even for him, "The amount of resources it was taking to understand and keep pace with each project was daunting."
The cloud grows up
As evident from Goldman Sachs' report, we're still in the earliest of days for cloud computing. In this early market, developers rule and, not surprisingly, developer-oriented infrastructure has won out. Developers have led to an incredible lead by AWS, one that only Microsoft's Azure seems to be able to cut into, given its foothold within the enterprise.
Those same developers drove OpenStack's community uptake, as well.
But that was the early cloud, and what worked for early adopters almost certainly won't be what drives long-term adoption, which is why both AWS and OpenStack need to figure out long-term growth plans.
Indeed, what Ted Chamberlin says of AWS is even more true of OpenStack: "AWS operates under the assumption that companies want à la carte functionality over intuitive solutions." Developers may, but the enterprises looking to the cloud for flexibility ultimately want as much complexity removed as possible.
According to one recent survey of European IT leaders, for "81% of European enterprises not using the cloud, the main stumbling block [is] insufficient knowledge of cloud computing." Asking them to comb through OpenStack's mailing lists and learn the intricacies of its virtualization-centric deployment model aren't winning arguments that will bring those cloud-shy enterprises into the fold.
AWS gets this and has aggressively rolled out enterprise functionality, according to Gartner's analysis, even as it tries to continually improve ease of use. OpenStack, on the other hand, keeps making the cloud harder to use. Until its community finds a leader that can solve this, OpenStack will simply be all community noise with little enterprise signal.