Your Microservices Effort Will Fail If You Only Focus On The Technical Details
22 Mar 2016
I have self-sensored stories about microservices, because I have felt the topic is as charged as linked data, REST, and some parts of the hypermedia discussion. Meaning there are many champions of the approach, who insist on telling me, and other folks how WRONG they are in the approach, as opposed to helping us work through our understanding of exactly what is microservices, and how to do well.
For me, when I come across tech layers that feel like this, they are something that is very tech saturated, with the usual cadre of tech bros leading the charge, often on behalf of a specific vendor, or specific set of vendor solutions. Even with this reality, I've read a number of very smart posts, and white papers on microservices in the last year, outlining various approaches to designing, engineering, and orchestrating your business, using the "microservices diet".
Much of what I read nails much of the technology in some very logical, and sensible ways--crafted by some people with mad skills when it comes to making sense of very large companies, and software ecosystems. Even with all of this lofty thinking, I'm seeing one common element missing in most of the microservices approaches I have digest--the human element.
I hear a lot of discussion about technical approaches to unwinding the bits and bytes of technical debt, but very little guidance for anyone on how to unwind the human aspect of all of this, acknowledging the years of business and political decisions that contributed to the technical debt. Its one thing to start breaking apart your databases, and systems, but its another thing to start decoupling how leadership invests in technology, the purchasing decisions that have been made, and the politics that surrounds these existing legacy systems.
I don't know about you, but every legacy system I've ever come across almost always has had a gatekeeper, an individual, or group of individuals who will fight you to the death to defend their budget, and what they know about tech (or do not know). I've encountered systems who have a dedicated budget, which only exists because the system is legacy, and with that gone, the money goes away too--sell me a microservices solution for this scenario!
Another dimension to this discussion, is that investors in microservice solutions are not interested in their money being used for this area. It just isn't sexy to be spending money on dealing with corporate politics, and unwinding years of piss poor business decisions, and educating and training the average business user around Internet technology. If you do not unwind these very human led, politically charged, business environments, you will never unwind the systems that exist within their tractor beams. Never. I'd care how much YOU believe.
In the end, I'm not trying to make you feel like you are going to fail. My goal is to encourage more investment in this area by the microservice pundits, vendors providing solutions the space, and VC's who are pouring money into these solutions. Many of the young, energetic folks at the helm of startups do not fully grasp the human side of corporate operations, and the potential quagmire that exists on the ground in front of them.
I am hoping that a handful of service providers out there can lower the rhetoric around their services and tooling, so that expectations get set at more realistic levels. Otherwise the push-back against the first couple failed waves of microservice implementations will become impenetrable, and blow any chance of making it work.