10 Things You Need to Know About the Hot DevOps Trend

by Chris Preimesberger

Based on a “2013 State of DevOps Report” by Puppet Labs, 66 percent of enterprises surveyed are already using or are planning on adopting a DevOps approach. The industry leading the adoption of DevOps is telecoms, with 88 percent of companies either using it or planning on using it.

High-performing DevOps organizations deploy code 30 times more frequently than traditional organizations. (Source: 2013 State of DevOps Report, Puppet Labs)

DevOps is a response to an extension of agile development practices. Agile development requires smaller, but more frequent code rollouts; operations staff must be working closely with the developers to prevent disruption.

Top operators are able to achieve this nirvana, but until you close the loop with operations, your continuous integration (CI) and continuous delivery (CD) are not really continuous.

Most legacy monitoring and management tools don’t fit in a DevOps environment. However, the tools that do fit are critical. Automation is fundamental to DevOps.

The top three skills needed for successful DevOps adoption are: a) knowledge of business priorities and strategies; b) knowledge of current business processes; and c) communication skills. DevOps does put a burden on Ops teams that are now expected to automate their environment. The mantra “Ops Must Learn to Code” is a frequent refrain—even if it is often unrealistic.

DevOps and cloud architectures are loosely coupled, change continuously, scale horizontally and are stateless. Legacy systems are unable to meet the latency, performance and scalability requirements that these new deployment models need.

If you are not treating your infrastructure as code, you are not doing DevOps. To do continuous integration (CI)/continuous delivery (CD) or continuous operations (CO), you must store configurations as code.

You can make your environment more stable by unleashing a Chaos Monkey (a test disaster incident). Disaster recovery plans are only as good as the last time you successfully used them. Doomsday scenarios need to be constantly tested.

Thanks to all the automation help, top DevOps operators have been shown to be much more productive than traditional IT. That’s one reason why IT workloads are migrating to SaaS and shadow IT, away from the control—and the budgets—of CIOs. However, enterprises can and are adopting DevOps, and fundamentally fixing IT as they do so, returning the CIO to the center of the discussion of how to improve the business.