While Amazon Web Services (AWS) may reject the private cloud, Eucalyptus' success may give it a way to have its public cloud cake and eat the private cloud too.
Eucalyptus was supposed to be dead. An early open-source cloud pioneer, Eucalyptus was one of the first to market with a private cloud but quickly lost traction as developers struggled to get it to work. Then OpenStack arrived to much fanfare (and even better open source bona fides), and Eucalyptus was doomed.
Except that it wasn't.
In fact, as I discovered this week in a conversation with Eucalyptus CEO Marten Mickos, the company is experiencing a significant renaissance with impressive product and customer traction, with one customer now running over 100,000 nodes. But more than one startup's near-death experience, Eucalyptus is a sign of a rapidly maturing cloud computing market, one that may well be dominated by public clouds like Amazon Web Services (AWS) but may still require a private cloud complement like Eucalyptus.
Not dead... yet
Eucalyptus was one of the first -- if not the first -- private cloud provider to launch. This was both a blessing and a curse. While it meant that Eucalyptus was pretty much the only private cloud game in town, until recently, no one has wanted anything but AWS.
But Eucalyptus wasn't merely early to the market. It was early with a product that largely didn't work.
I worked at Canonical in 2010, and we promised to be one of Eucalyptus' big, early partners. Unfortunately, we struggled for months to get it to work and ultimately gave up to adopt OpenStack. We weren't alone in our frustrations.
While there were companies that ran Eucalyptus successfully in significant production, most would-be buyers lost interest in undertaking the challenge when OpenStack hit. Despite Eucalyptus investing heavily to improve its product in 2011, the market was gaga for OpenStack.
Eucalyptus and its private cloud dream seemed dead.
New life into an old project
Three years later, however, the tide has turned in Eucalyptus' favor. Talking with Mickos over lunch at an event in Vail, Colorado, he offered four reasons for Eucalyptus' rise:
- Eucalyptus works great.
- The team is cohesive and very strong.
- The market finally recognizes Eucalyptus' value proposition.
- People increasingly recognize that OpenStack is not the answer.
On the first point, Eucalyptus has been solving one of the biggest headaches for those running apps in public clouds: management. By building out its operational tooling, as well as core infrastructure improvements, Eucalyptus works, and it works really well. At last.
As to the third point, Mickos told me that most companies start with AWS and then determine they need to complement public cloud resources with private cloud resources to get better performance, more control of their data, and lower costs. Increasingly, they turn to Eucalyptus for this because it offers semantic and API compatibility for nearly 10 AWS services.
The OpenStack math/myth
And then there's OpenStack, which Mickos says customers find "too complicated." Ironically, and similar to the early days of Eucalyptus, Mickos says customers also find that OpenStack doesn't work.
This is, of course, heresy. OpenStack is the darling of the open-source counterinsurgency against AWS. One glance at Google Trends (Figure A) and it's clear who is winning, right?
Google Trends results of OpenStack and Eucalyptus.
Or how about jobs advertising the need for OpenStack or Eucalyptus skills (Figure B)?
Job trends of OpenStack and Eucalyptus.
These don't tell the full tale. In fact, Mickos notes on his blog that Eucalyptus just closed a deal with an OpenStack contributor. As he told me over lunch, the company in question is one of OpenStack's largest contributors.
OpenStack's problem, he stresses, is not one of community -- but rather, one of committee:
"OpenStack... is also facing very serious challenges not unlike those that Unix and CORBA went through in their days. It's difficult to produce technically brilliant products when governance is shared among very large corporations, each one with their own agenda. In an all-embracing collective it is difficult to say no to new ideas, but 'no' is a vital component of designs that win."
I agree, and it's why I've argued that leadership from Red Hat bodes well for OpenStack's ability to streamline its product and compete. But for now, Eucalyptus' improved product and AWS compatibility are significant selling points.
Does the private cloud have legs?
Which brings us back to the private cloud. While most signs point to AWS dominance, the resurgence of Eucalyptus suggests that there is ample room for private clouds to grow alongside (and, indeed, out of) AWS. Even though AWS is fully committed to public clouds, it will not build these private cloud resources.
Meanwhile, Microsoft Azure, in particular, poses a real threat to AWS precisely because it marries public cloud services with data center assets. AWS may not want to touch the private cloud, but it benefits significantly from having the Eucalyptus suckerfish giving its customers an option of running hybrid workloads.
Given that some of AWS' constraints aren't likely to get better anytime soon, according to investor Brad Feld, who argues that "some of the things [AWS] need[s] to be doing to maintain their dominance is just not in their nature," Eucalyptus may have a bright future as a hybrid cloud "suckerfish" complement to AWS.