Your Microservices Effort Will Fail Because You Will Never Decouple Your
I’m regularly surprised by companies who are doing microservices which are failing to see the need the change organizational culture, and that microservices will be some magic voodoo to fix all their legacy technical debt. That simply decoupling and breaking down the technology, without any re-evaluation of the business and politics behind, will fix everything, and set the company, organizations, institution, or government agency on a more positive trajectory. In coming years, we will continue to hear stories about why microservices do not work, from endless waves or groups who were unable to do the hard work to decouple, and reorganize the operations behind the services they provide.
The monolith legacy systems I’m seeing targeted are widely seen as purely technology, which is why it is often labeled as technical debt. What is missing from this targeting and labeling is any acknowledgement of the people and decisions behind the monolith. The years of business, political, and cultural investment into the monolith. How will we every unwind, or properly address the monolith, if we do not see the organizational, human, and business aspects of why it exists in the first place? Are we talking about the business decisions that went into creating and perpetuating the monolith? It is highly likely we will be making some of the same decisions with microservices, which could end up being worse than when we made them with a single system. Distributed mess, is often more painful than consolidated mess.
I’m seeing endless waves of large organizations mandating that their teams invest in microservices, with no mandating for microteams, microbudgets, microdecisionmaking, or any of the other decoupling needed to make microservices truly work independently. I attach micro as a joke. I really don’t feel micro is the constant that needs applying when it comes to services, or the business and organizational mechanism behind them. However, it is the word du jour, and one that gets at some of the illnesses our organizations are facing in 2018. In reality, it is more about decoupling and decomposing the technology, business, and politics of our operations, into meaningful units that can be deployed, operate, and deprecated independent of each other.
My point is that your microservices effort will fail if you aren’t addressing the business side of the equation. If your microservices team(s) still exist within your legacy organizational structure, you really haven’t decoupled or decomposed anything. The old way of making decisions, dealing with budget impacts, will still reflect what happened with the previous monolith. Your technology will be independently operating, but still beholden to the same ways of deciding and funding what actually happens on the ground. The result will resemble having a entirely new motor, where you are running without lubricant, or possibly old, thick, expired lubricant that prevents your new motor from ever delivering at full capacity, and eventually breaking down in ways you have never imagined while operating your existing monolith.