Like other software that originated in academia or government labs and ended up an open source project, OpenStack was sufficiently complex, incomplete and unfined to scare all but the most sophisticated IT organizations away from what promised to be a substantial DIY implementation project. There's a reason Microsoft, Oracle and VMware software dominate corporate data centers despite their cost and often infuriating licensing terms: IT execs know they'll get a complete product which runs on existing systems and backed by a deep support bench from a vendor they know and (mostly) trust.
Still, if Linux proved anything, it's that one should never underestimate the potential for thousands of developers to slowly grind away at bugs, add features, perfect and simplify installation processes and tighten the security of open source software. Remember, the tortoise wins in the end, and in software development, the cacophonous bazaar often ends up building a better product than a few high priests in the cathedral. For this reason, it was always unwise to make long-term bets against OpenStack's success.
However, those of us that watched its early growing pains in light of historical flops of paradigm-shifting efforts like ATM (the networking kind), Openoffice and the semantic web (3.0) can be forgiven our skepticism.
After a typically turbulent trip through the Hype Cycle, OpenStack appears to be clawing out of the depths of disillusionment on its way to productive utility within many enterprise data centers. Indeed, this time OpenStack advocates have more than just isolated anecdotes to crow about, rather there are credible financial numbers showing corporate commitments to private OpenStack clouds.
OpenStack fueling growth at Red Hat
Red Hat has been on a roll for years with revenue double what it was four years ago and operating income up about 50% as it rides the wave of interest in open source software to become the dominant supplier to enterprise IT. It's latest quarter was no exception, with revenue up 19% and income 7% higher year-over-year, however what stood out was the source of Red Hat's growth. Sales of its application development and emerging technologies (read, OpenStack and OpenShift container-based PaaS) rose 33%, while large enterprise deals exploded. According to CFO Frank Calderoni, Red Hat closed a record number of million-dollar deals in its second quarter ended August 31st, up more than 60% year over year. During the company's earnings call, Calderoni added more details,
"Looking sequentially at our top 30 largest deals metric, all of the deals were $1 million or more and five of those deals were in excess of $5 million. Cross selling was strong, with over 70% of the top 30 deals including one or more components from one group of Application Development and emerging technologies offering."
However what got my attention was later in the call when CEO Jim Whitehurst summed up the key to Red Hat's growth this way [emphasis added]
I think there are several factors that are all coming together. If you look across those deals, we're doing I think an even better job of selling the portfolios. So seven of the top 30 deals had OpenStack in there, nine had RHEV (Red Hat Enterprise Virtualization, a VMware competitor based on Linux KVM). …
"If you look now, we have over 1,200 customers buying one or more of our emerging technologies. We had three OpenStack deals alone that were over $1 million. So I think we're seeing really, really, really good traction there.
OpenStack is certainly no longer a niche product when nearly 25% Red Hat's largest deals include it and 43% of those are in 7-figures, but that's just for starters. Whitehurst said that its customers are still early in their private cloud adoption, however the sales process is more complex than for its OS products, and thus filling the order pipeline takes time,
Without a doubt, selling OpenShift, OpenStack is a much longer effort to sell those things and so those are higher cost in the short run. But that’s [sic, let's] recognize the difference between the subscription and the lag on revenues versus the activities we're doing now to sell things for the future.
Enterprise OpenStack also received a lot of attention at Veritas Vision (see my analysis of the event and the company's strategy) where Veritas announced a partnership with Red Hat to build storage, data protection and workload automation technologies for enterprise private clouds. First out is HyperScale for OpenStack providing policy-based workload deployment and live migrations and performance improvements by offloading backups and snapshots from application hosts. Indeed, enterprise interest in OpenStack is high if my experience of attendance at several OpenStack sessions at Vision is any indication, as the rooms were full of people with plenty of questions.
Rackspace results also reflect OpenStack adoption
Although its growth isn't as robust, OpenStack pioneer Rackspace also reported a 2nd strong quarter in August in part due to OpenStack adoption. While the company didn't specify sales figures, Rackspace executives did highlight OpenStack in their earnings call. Again noting the long OpenStack, President and CEO Taylor Rhodes said,
We've seen strong demand for OpenStack private clouds. In Q2, we closed deals to deploy OpenStack private clouds for a leading automaker and for one of the world's largest IT and management consulting companies. We also saw a major expansion of our relationship with a very large customer that we landed in Q1, which is one of the world's leading industrial conglomerates. The customers for our OpenStack private cloud tend to need larger more complex solutions and these have a longer deployment and revenue materialization cycle averaging about six months.
In response to a question, Rhodes later added,
In terms of OpenStack Everywhere, look, I would say that we're experiencing a fairly good inflection of demand in conversation and pipeline activity from that. I think it's common sense to understand why we think generally there is a large base of market out there that is going to want private IS [infrastructure services] for some time to come.
The deal expansion Rhodes mentions is for one of the customers he highlighted in Rackspace's Q1 earnings call when he said [emphasis added],
In Q1, we closed deals to deploy OpenStack private clouds for one of the world's leading manufacturing conglomerates and for a global bank based in Europe. Our OpenStack private cloud set a new record for bookings in Q1. Revenue is growing in the high-double-digits and our pipeline is growing even faster.
Admittedly it's a small base, but that kind of revenue growth is in Amazon Web Services territory.
Despite my belief that in the long-term superiority of the public cloud business model providing infrastructure and software as a utility service, the transition from private data centers to public utilities won't be an overnight discontinuity, but rather a generational migration. As Rhodes admits, private clouds
...may be a transitionary platform that in five years to 10 years, doesn't exist anymore, ... but we do know that companies, large and medium particularly in size and in key verticals like financial services and media and entertainment are looking at private IS as part of the overall environment.
VMware has long been trying to exploit its enterprise dominance in virtual servers by expanding its portfolio into a full-fledged cloud, but as I wrote after VMworld, the company will remain shackled to its legacy for quite a while. Microsoft is also aggressively pursuing private cloud infrastructure and once Azure Stack matures into a complete product, will have the most compelling hybrid architecture, at least for organizations not already wedded to AWS.
OpenStack represents an interesting middle ground: pure private cloud technology, unencumbered by legacy product baggage. Now that vendors like Red Hat, Rackspace, IBM, Veritas and others are imbuing the technology with enterprise features, reliability and support services, OpenStack is a viable alternative cloud platform. Indeed, rather than continuing to write its obituary, look for more vendors, analysts and customers to embrace OpenStack in the coming year and if it follows the path of Linux, OpenStack will further disrupt the cloud vendor landscape.