Devops at Adzerk

added by bsenoff
10/19/2011 3:46:32 PM

276 Views

From our sister company, Adzerk, James Jeffers speaks about how the company has designed it's software process to get around the dev/production bottleneck inherent in most companies.


3 comments

dpeterson
10/19/2011 4:00:02 PM
The company I work for in my 9-5 job follows a similar approach, and has since their inception 13 years ago. We don't follow a true continuous integration strategy, but our code-base is living and it's not uncommon for half a dozen builds of any one of our applications to go out in a day. We are very much agile without all the pomp and circumstance.
<br>While sometimes this does lead to problems: poor QA, botched deployments, botched database schema updates; it also gives us some pretty sweet benefits. We've been able to stay lite on staff for many years, currently we employ 10 full time developers, 2 of which are founders of the company who still code daily. Every single developer is expected and prepared to deploy code to any one of our clients, depending on the situation. There's no waiting for so-and-so to get back from lunch or sick leave or vacation. We just signed on our 80th client, and we're not talking retail software here. Each client setup involves a full gap analysis and design document (for any changes to our base software), server and client machine setup, deployment, training, and full support with no restrictions on hours over the life of the contract. We can do all that with fewer than 30 full time employees, and it is largely attributable to not having a dedicated ops team.

vijayst
10/19/2011 8:54:48 PM
If every developer takes care of the operations, then it is a devops team, right? The whole idea is not to have a dedicated operations team and back-and-forth communication going on between these two teams for bug fixes and enhancements.

I used to work for a company where the product management takes the role of prioritizing what should be done. So, we have another team to coordinate with.

dpeterson
10/20/2011 12:06:47 AM
That's a typo. I meant to say "largely attributable to not having a dedicated ops team".