The Chef Win Series is a response to: "How has Chef changed our operations for the better, in the past 2 years since our shop adopted it?"   Example:

Obnosis: Paradigm Shift

Please don't blame us for the fancy "Paradigm Shift" terminology, it's a side effect of work by others. 

  • Continuous Development - the trend
  • DevOps - the trend

Jeff Bezos of Amazon fame suggests to "ask yourself what won't change" is the key to markets, and what people could never do up to now is get past the quagmire that is deployment in the enterprise. That just wouldn't change.

For example, one shop in 2005 or so where a relatively un-impressive app was given to a 4 man team 6 months to work out a tested deployment of a release. The entire team spent the entire 6 months getting that worked out, only to repeat again on the next cycle. This was worse than, but only barely, the previous contract in 2003 at a different big shop with a similar configuration.

With that kind of cost nightmare, no wonder enterprises wanted to pack in as much into a release as they could.

Now it's an automated pipeline, or at least it is in the better shops. A code change and a new tested release can be the same thing, if you wish them to be. But that's such a huge change, we can't even imagine all the downstream consequences.

Chef and Micro-Services: the Effect

Rather than looking at micro-services as a goal or strategy, consider it as a lucky side effect of Continuos Delivery and Devops. When everything is automated and easy, now it's possible to think in terms of splitting up an asynchronous app into micro-services. Are there benefits?

There are both benefits and new unexpected consequences to splitting an app up into micro-services. We won't discuss them here, but if you're a software vet, you can probably think of a few of each.

What is important, as discussed in this Thoughtworks article on MicroServices, is that we won't know about all the effects of micro-services on our industry for a few more years.

What Role Does Chef Play in this Dance?

Chef enables experimentation. You can't experiment when it takes a long time to switch things around - nobody has the resources and patience. With Chef, you simply change up your deployment scripts.

There will be changes in the marketplace when all this settles out. Nobody knows what those changes are. But Chef will have played a key role in bringing them to the fore.

What's Your Bet?

Do you have a bet on which way this goes?

My bet is that micro-services disappears as a phrase, but instead becomes "the way" that we do things, in a very few years. We won't call it that, we'll just be doing it. 

When the shift happens, that will cause a few other shifts, especially economic shifts. Because risk mitigation is entirely different inside small pieces than large pieces. Economics will change the sourcing model because risks themselves have changed. Outsourcing and local sourcing will compete in entirely different ways than they do now.

Lots of fun changes coming down the pike.