For this, admittedly first, edition of Been Here: Done That, I’m discussing my thoughts and experiences of the DevOps Enterprise Summit 2016. The conference occurred over three days, Nov 7th to 9th, of 2016. Main track, side tracks, vendor hall, all the standard stuff. When I heard about this conference, the thing that kept my attention was the word “Enterprise.” I know DevOps can scale down. Automation, cross-function, transparency; these are three things essential to getting things done in small teams. I was curious to see what this culture looks like when you have a whole lot of scale to throw around.
Straight off, the size of the presenters’ organizations screamed “Enterprise!" Never mind that, the size of the IT Departments of these organizations was astounding: some in the four to six thousand range. Three zeroes. Ten cubed. You be Steve, I’ll be Frank: If a set of practices will scale to that level and provide even single-digit improvement in revenue, we’d all better be paying attention.
The most successful presentations, from my perspective, were the ones that overviewed the high-level strategies an organization was employing, but then dug down into one or two efforts in order to prove a point. For example, Topo Pal of CapitolOne talked about the big vision of their multi-year transformation, but nailed home his view of an Automated Pipeline. That nail takes sixteen swings to drive in. Sixteen aspects that tell you if your pipeline is whole. The 0x10 Commandments.
DOES16 is sending me away with a wealth of shared experience, and a new book. I didn’t mention that, before. Gene gave away his new book. It’s like he knows that I’ve got a long layover and an insatiable hunger for work transformation techniques. That is, if I can stop reading ‘Dune’ for any length of time. I’m starting to see DevOps in my SciFi.
We should all be heartened by the fact that these transformations were either multi-year in duration, or were still underway. New ideas need relentless experimentation to take hold, and they must be content with doing so at a grassroots level for long periods of time. The initial target then, should be to produce data. Data which establishes the alpha state. Data which will be referenced when we ask ourselves if our experiments made things better or not. After all, how can you be part of a success story if it has no beginning?
Topo referred to them as "the Ten Commandments, in hex.” I’m still mostly directly quoting him, but changing the notation on preference.