Robbins has unimpeachable credentials when it comes to cloud computing. He ran the world's largest online retail operation, Amazon.com. His company Opscode's open source "scripting on steroids" tools, known as Chef, power RightScale, Rackspace and Engine Yard. They're also heavily used by VMware. Clearly, if you want to run a cloud, Chef is a good place to start. Opscode has also just launched the Chef Platform as a Service (PaaS). But he's not the only one with cloud credentials.
Chris Pinkham was head of the team that invented Amazon EC2 and blazed the way for a commercially successful cloud computing environment. He pioneered ideas about data center operation, scale, and end-user interaction that reshaped information technology. His startup, Nimbula (formerly known as Benguela), is also aimed at enterprise IT departments.
The connections go deeper as well: Pinkham's original collaborator on EC2, Chris Brown, is now the VP of engineering for Opscode. And the third member of the original EC2 team, Willem van Biljon, has joined Pinkham at Nimbula.
In what some would call a bellwether of where the IT industry and the cloud computing market are heading, both Amazon alums are aimed squarely at delivering the promise and the functionality of the cloud within the enterprise, but by different paths.
How the two companies will tackle the enterprise
Opscode's Chef Platform is a management portal that allows users to use Chef as a service rather than maintaining it internally. Users get an interface that lets them pick and customize automation scripts, known as recipes, and run them against their own server environments or external cloud services.
Meanwhile, Nimbula provides cloud operating system software that is self-aware, automated and can scale to a massive degree on commodity hardware. Linux-based, it operates from a central network boot image and is basically an automated, high-functioning hypervisor and virtual networking platform. It deploys onto whatever server resources you point it at, provides a REST application programming interface (API) and Web portal that can be used to launch and manage virtual machines (VMs) and can manage its own networking and storage.
Pinkham said Nimbula's software is "well integrated" into existing enterprise IT governance and management structures. He also claimed it is a step forward over other cloud platforms like Eucalyptus in the degree to which it scales, which he called "unprecedented" for such a complex system. Pinkham said Nimbula suffers little to no performance attrition even at a very large scale, 100,000 nodes or more.
That kind of talk is easy to discount from a software vendor; it's a pitch, after all. But given Pinkham's background in creating the Amazon cloud, it's harder to dismiss.
After all, in 2007, EC2 left many scratching their heads and wondering how Amazon was going to be able to make money on hosting, no matter how nifty the self-service aspect seemed to be. To compete in a saturated hosting market and bill by the hour would take an operation that had an "unprecedented" level of automation and ability to scale (and remain profitable)…so if anyone can make a credible claim these days, it would be Pinkham.
Nimbula has already signed up a financial services firm with a long-term plan to convert its compute farm into a private cloud environment. The customer declined to be named. Another, Metropolitan Health Group, said Nimbula was the best of several cloud platforms it had evaluated for internal use but didn't say whether it was moving forward with plans for implementation.
Pinkham said that he wasn't pitching enterprises on getting rid of their current operations and replacing them with his cloud operating system (OS). He sees cloud computing as a gradual shift that will eventually make operations almost irrelevant to creating and performing computational tasks.
Opscode's plan for enterprise dominance
Opscode's Robbins shares Pinkham's views on the thirst for cloud within the enterprise, but differs on the best way to achieve it. Robbins wants to help admins automate what they already have.
Like Pinkham, he was careful to say that existing operations didn't need to switch to operating like a cloud, but would want to in future.
"I'm not going to pitch them on the cloud; if it's working for them, leave it in the back office," Robbins said.
He said the point of Chef was to help get rid of mundane tasks and let operators focus on better service delivery.
"Configuration management is an interesting and sort of arbitrary set of CompSci problems that most people don't care about," said Robbins.
He says virtualization is widespread at this point, even though it took four or five years to really get going. Similarly, he said cloud would take a while to sink in, but in the meantime he can help administrators and developers take the edge off their cloud hunger without undue commitment with the Chef Platform and automation tools.
Can the ex-Amazoners overcome the competition?
Both companies are rolling in clover, too. Opscode landed $11 million in venture capital this year, and Nimbula has $6 million to play with. It's an interesting dichotomy, almost as if EC2's early history is playing out all over again, but this time the target is not the lone developer but the enterprise data center.
Despite their pedigree, Opscode and Nimbula face a slew of competition, especially in the enterprise market. Aside from startups like Eucalyptus, Abiquo and Enomaly (which Pinkham characterized as good efforts but not truly sufficient), really large firms like CA and BMC are making every effort to turn enterprise customers on to cloud technologies. They're not going to have to fight for a customer base, either; they've already got one.
Oracle and IBM claim to deliver private cloud at every level, and cloud and virtualization management and provisioning tools litter the landscape, from startups like RightScale and enStratus to firms like VizionCore. VMware, too, has said it will deliver the private cloud in short order.
Pinkham and Robbins will need a compelling argument to rise above the fray. One thing is clear, however. There is plenty of room for both to try to put the lessons of Amazon.com into practice as the gradual, inevitable shift towards cloud in the enterprise begins. Both Robbins and Pinkham say that the whole idea is barely off the launch pad.
"The promise of cloud computing is as yet unmet," said Pinkham.
Carl Brooks is the Technology Writer for SearchCloudComputing.com. Contact him at firstname.lastname@example.org.