Jesse Robbins (@jesserobbins) is the co-founder of the Velocity Conference and co-founder and chief community officer of OpsCode, the company behind the popular infrastructure automation tool Chef. During a recent interview, we discussed the current state of cloud infrastructure automation, the composition of the audience, the shifting roles for both developers and sysadmins, and an example of a large-scale use of Chef by Cycle Computing.
Highlights from interview:
On the evolution of the sysadmin
“What has happened is that the old school operator, where you had a three-ring binder and you followed procedures over and over again, frankly, that job sucked. What has happened over a period of time is that sysadmin skill set has been expanding to include more and more basic infrastructure software development. What has been most interesting in that transition, has been essentially, I think of it as an expansion in the job roles going from basic systems administration to really what we call infrastructure engineering. The road ahead for everybody, who builds and maintains infrastructure or applications in software, looks like building a very powerful software platform.” [Discussed 01:47]
How much awareness is there of technologies like Chef, Puppet, and CFEngine?
“When I look at this and I think, ‘Where are we?’ we are probably at less than 1% of the total addressable market. We have got a long way to go. It’s a big space. It is in giant flux. We are moving from software developers, sysadmins and their management up to the CIO and CEO. That’s going to be a decade-long revolution, just like it has been with open source.” [Discussed 04:40]
What’s the most surprising use of Chef so far?
“Large-scale stuff … for instance, the work that Cycle Computing is doing taking the Amazon spot market for large instances. They are combining that with a tool they wrote called Grill, which is based on Chef.
“They are running these 50,000-core protein folding jobs to find cures for cancer. They are building a supercomputing cluster in a couple of minutes, where the consumer of that is the scientist. It’s not like they’ve got some big engineering team. It’s almost like a single click or a few clicks and these very, very large, typically previously prohibitively expensive workloads. They are using Chef for that.
“It’s unbelievable. 50,000 cores in a couple of minutes for under $5,000 an hour, where previously that would have been a massive, multimillion-dollar, multi-year supercomputing project. That stuff is available on demand.” [Discussed 06:53]