Much has been written about the emerging trend of DevOps. Mike Gualtieri, as always, provided sharp commentary in his blog titled “I Don’t Want DevOps. I Want NoOps.” I was very excited to attend IBM Pulse 2011, which is aimed clearly at operations professionals and is Tivoli’s primary event. Over the past year, the Tivoli brand has gone through a management change. Danny Sabbah and a lot of his management team moved from Rational, the developer brand, to Tivoli. So I was interested to hear how Danny Sabbah and his team had introduced the ideas of Jazz and OSLC into the Tivoli mission. And I was not disappointed. The keynotes, track sessions, and analyst and press events were all aimed at vision. They talked about the end-to-end value chain of delivering and maintaining software. They placed heavy emphasis on process and automation, describing a vision of integrated tools and process allowing software to move seamlessly from idea to implementation and beyond. IBM is not alone in this vision; increasingly, other vendors are trying to break down the barriers between operations and development. HP, with its Business Technology Optimization (BTO) vision, is connecting quality, portfolio management, and service management. Microsoft for a long time has described an integrated vision for the .NET platform.
BUT, was the audience of Pulse 2011 interested in DevOps? Perhaps the best barometer was the Service Design Delivery track. This track was all about DevOps, titled Agile Operations by IBM. It was staffed by IBMers from the Rational and the Tivoli brands and included a number of customers who were integrating their delivery and operations tools and processes. The track content was strong, but unlike the baseball field on Field of Dreams, IBM built it and they did not come. The track was poorly attended. There was some interest in the ideas of Agile and integration, but compared to the other tracks, such as Service Delivery and Automation, attendance was low. This highlights the fundamental issue with the DevOps movement. Who has the problem that DevOps fixes? Often development and operation professionals are measured within their own domains. Neither party is measured on the whole value chain. Development professionals are measured on project metrics such as time, budget, and quality. Operations pros are rewarded on service measures such as up time and cost. Both parties want a better life and to deliver more business value, but without a clear shared objective, it is difficult for them to get motivated. For vendors the problem presents itself in finding the buyer. Often the only place these two camps come together is at the CIO level, a level that is already busy with numerous change programs.
So is all lost for DevOps? No, it will happen, and not just because of the efforts of vendors but because the movement is rooted in business need. The business needs software to be delivered faster and to provide increased value. Optimizing development or operations will not give if these groups are treated in isolation. My observation is that:
- DevOps will become the reality in situations where the business and technology require it. Thus mobile and new web apps will motivate not only different development models but also the transition to DevOps.
- Development will push the agenda. Agile methods have been the catalyst for DevOps in many organizations, and as Agile continues to grow in use, the drive to DevOps will happen.
- Organizations will adopt technologies such as cloud, making traditional operational processes obsolete.
I would love to hear what you think. Are you deaf to DevOps?