DevOps Innovation

DevOps in all its forms is really taking off—whatever ‘DevOps’ actually means. Well, for me, it’s continuous, agile, pipelined delivery of automated business services, controlled by end-to-end feedback loops—and ‘taking off’ means that the big boys are buying up the capabilities needed for ‘DevOps at Scale’.

An example is the purchase of UrbanCode by IBM; and (since not everyone engaged in Agile likes big companies) the risk is that we get into religious wars about whether DevOps at Scale is really, properly, Agile—which are just a bit pointless. DevOps is about bringing increased agility to the delivery of automated business systems and it’s a journey, not a state of religious grace. A more sensible risk to worry about is whether DevOps innovation can survive big company politics, even if an acquisition keeps all the key people that built the acquired product (which is probably why IBM is maintaining a separate UrbanCode culture.

This all means that some people will doubtless still want to follow the remaining independent DevOps vendors such as Electric Cloud. Personally, I think that choice has its place, but that cultural issues in a company—whether it can cope with Agile practices, can overcome existing silo boundaries, can deal with the political impacts of continuous delivery—are often going to be as, or more, important than its choice of vendor; but talking to Jim Ensell (Chief Marketing Officer at Electric Cloud) and Rohit Jainendra (its Chief Product Officer) has been very interesting nonetheless. Electric Cloud has some excellent technology, especially around optimised use of cloud resources for rapid software builds, and some attractive innovations. It is also a privately-held company, which (given that its strategic vision is sound) means that it should be able to concentrate on its customers and their long-term goals, without short-term distractions from the stock market.

Electric Cloud is obviously aware of the current activity in the DevOps marketplace and has revamped its web presence. It is also co-sponsoring (with IT Revolution founder Gene Kim, co-author of my favorite DevOps—and Business Continuity—book, the Phoenix Project) the DevOps Enterprise Summit (Oct 21st–23rd 2014) in San Fancisco. This is all good; too often private companies are so focussed on their vision and their strategy that they forget to tell anyone about their existence.

Electric Cloud’s recent innovations are ElectricFlow and ElectricAccelerator Huddle. Both seem to address the new demands of continuous delivery at scale—DevOps has grown up. Electric Flow is a new family of applications targeting Enterprise needs for standardised software deployment; Electric Cloud is particularly proud of the way it satisfies DevOps practitioner’s needs, and of its clean UI.

Electric Cloud’s Huddle (no, not the other Huddle) is in some ways even more interesting. It lets a team become more productive by tapping into unused desktop CPU capacity within a peer-to-peer group of developers, in order to make continuous delivery faster. It could, in my opinion, increase agility in big organisations by allowing teams to take control of their own destiny without impacting corporate technology strategies (oh, and ElectricAccelerator Huddle is free)—and ElectricFlow should provide the management insight and controls needed for ‘just enough’ governance, if appropriate. The existence of two Huddles, however,  is potentially more confusing than Electric Cloud thinks, in my opinion; a team using Electric Cloud’s Huddle for collaboration around build acceleration might well use the other Huddle for collaboration with business users and the breaking-down of silos.

It seems to me that Electric Cloud can tell a good story around starting small and growing into Enterprise-scale continuous delivery without discontinuities. This is important as ‘the next Facebook’ is probably a 2 person start-up with tools that can’t scale, which it will use well past the stage when it needs something better. In the chaos as it grows past the point where its tools really can’t cope, it may die. On the other hand, if a 2 person company invests in a typical enterprise-scale scale tool at the beginning, it will probably go broke before it gets started. The better enterprise-scale companies are now realising the importance of low cost-of-entry options (often with SaaS solutions) and seamless growth.

Another subject that came up in the discussion with Electric Cloud was the possibility of extending its accelerator technology beyond build automation. Test automation may be Electric Cloud’s next target; and gaming companies are expressing an interest in OEMing it for the acceleration of rendering. So, Electric Cloud looks to have a good future, on top of its current innovations.