Mickos: Amazon and Eucalyptus still rule the cloud
Ex-MySQL man takes fight to OpenStack
Open source or open core?
Open-core became a talking point when NASA dropped Eucalyptus and went with OpenStack. According to NASA, it wanted changes made to the code, but Eucalyptus was unwilling to make them. Mickos tells a different tale. He says the sticking point was that NASA had a problem signing away copyright and wanted changes to the Eucalyptus contributor agreement before it would submit code. Mickos says Eucalyptus made changes to its contributor agreement, but then NASA went silent and didn't follow up.
"When somebody contributes to us, we require them to sign over copyright, but so does the FSF [Free Software Foundation] and Ubuntu. There's nothing new in that. If somebody contributes something significant, we pay them money so they can make real money.
"We don't see it as an issue there and we think it's important to have the copyright well managed. It's a part of our governance model. We've told the world this is how we operate, we are in charge the roadmap and we require copyright sign over when people contribute," Mickos said.
Mickos hints at a more prosaic reason for the breakdown between Eucalyptus and NASA. That Eucalyptus was still a relatively small technology-focused shop of just 15 engineers, unprepared to handle something like customer service. Today, Eucalyptus has 55 staff, and meetings every Monday open with an update on any customer's issues, bugs, and bug fixes. "Maybe we didn't remind them, we didn't reach out and call out and say: 'Hey guys where are you?' because we were small then," Mickos said. "We have made sure as an organization we are more out reaching."
Open core might have bothered the boffins inside NASA, but it has not stopped others from downloading the software or even paying Eucalyptus for those more enterprise-y features.
Mickos won't say how may paying customers Eucalyptus has, but he claims there are 25,000 Eucalyptus clouds, and he says you should divide this by 100 to get the number of paying customers. This is a classic smoke and mirrors tactic from open-source companies, who like to talk uptake rather than paying customers. The company is growing, but it's not making a profit, Mickos admits.
Those people using Eucalyptus are running it on thousands of hardware servers, and this means it's on a larger number of virtualized servers. The latest customer is European social network gaming site Plinga, who Mickos claimed runs hundreds of servers and is doubling its server footprint every six months.
Recently, the company's bruised ego received another fillip. Eucalyptus was invited to join the anti-VMware, pro-KVM hypervisor party started by IBM and Red Hat: the Open Virtualization Alliance, also announced this week at OSBC. So far, Eucalyptus is the only pure cloud player on-board, although the Alliance says that more companies can't wait to join up.
"They invited us to this because they know we have the widest installed based of all cloud platforms," Mickos said.
We don't know who said what to who or how it really went down, but there is a pay-off for Eucalyptus. The Alliance is a big marketing push for KVM against VMware by companies with cash and contacts: IBM, Intel, Hewlett-Packard. Both may help Eucalyptus gain entree into new, bigger customers in the enterprise, gaming, mobile, and government markets.
Mickos just spent twelve months getting Eucalyptus into shape. The company has improved its technology and documentation, begun a professional services operation, and has a version 3.0 due that uses algorithms to continuously check the health of an Amazon cloud and run failover should it spot a problem. "We are on a completely different level," Mickos says.
But for all this work, more lies ahead if Mickos and Co. are to make it big. While Amazon still dominates the cloud, the web's honeymoon with the service is over, and devs have more choices of technology and development model than before.
Mikos, who helped build his former venture MySQL into something considered worth $1bn, is philosophical but positive. "Building a big successful business is not about one step, but about taking many useful steps and getting a small victory here and a small victory there, and that’s what adds up," he says. ®