Fred George talked about what organizations can do to successfully deploy microservices at the GOTO Berlin 2015 conference. InfoQ interviewed him about business and IT interaction for microservices, what organizations can do to support teams in using microservices, benefits of microservices and what the future will bring for microservices.
InfoQ: Can you explain how IT can support the business when organizations are adopting microservices?
George: Microservices enable rapid delivery of new business ideas. It allows the business to move from hard requirements (traditionally signed in blood) to trying ideas in the minimum time at moderate cost. If the business isn’t ready to deliver quickly, much of the value to the business of microservices will not be realized. It is a fact that IT trained the business to commit to requirements up-front over the last few decades. We (IT) need to take the lead in reversing this. The situation is compounded by explicit departments vested in managing to these firm requirements. Such groups will actually hinder exploitation of microservices .
InfoQ: How should business and IT interact if they want to truly exploit microservices?
George: There is a key paradigm shift required, but only for certain types of problems. We want to move from "requirements" (or stories or whatever else you would like to call them) to "trying out ideas". This is particularly required for "fuzzy" problems where precise answers are not forthcoming. These are the domains of the recommendation engines: Day trading; Google advertising; Amazon book recommendations; Questions like "should I lend you money?" These fuzzy problems lend themselves to continued refinement of the algorithms. Some ideas will be spectacular successes; some will be abject failures; most will have little or no impact. The key here is to try an idea, measure the results, and react accordingly. Then repeat, day after day.
The business is unlikely to be ready to interact at that level. Indeed, the concept of trying something and failing may be considered professional suicide. That needs to change. Once that business attitude shifts, the business interacts with IT differently: The business explains the core concepts and identifies the KPI’s; with that understanding, IT starts trying ideas with the coaching of the business. The resulting relationship is much more intimate and continuous. Indeed, the traditional business analyst role is a hindrance.
InfoQ: Do you have some examples of successful Business-IT interaction? Which benefits are they getting from it?
George: The seminal work done at Forward Technologies in London is a great example. Originally, they worked in Internet advertising, a fuzzy problem. The developers were taught the domain and identified the KPI’s. Then they started experimenting. The developers’ ability to react quickly provided significant competitive advantage. They employed the same concepts in other domains: e-commerce in pet products, and energy contract switching via uSwitch. Their model of working was copied by Compare the Market, also exploiting MicroServices. BTW, Forward had neither BA’s or QA’s!
There are also some examples in the financial trading environment, another environment where rapid experimenting yields superior profits.
InfoQ: What can organizations do to support teams in exploiting microservices?
George: The most important thing for the organization to do is to trust the team. As in any transformation (organizational or technical), things will go wrong. The naysayers will pounce on every little failure as a failure of the philosophy. The naysayers are ignoring the benefits of rapid delivery. They don’t believe that "experimentation drives innovation", the mantra for Forward. They demand perfect deliverables, a fiction in the big-deliverables world. At GOTO Berlin, Dr. Nicole Forsgren reported on her multi-year study of rapid delivery. She reported reduced risk in rapid delivery to the business.
Secondary, but also key, is support of rapid delivery in all supporting processes. This would include using cloud computing, continuing interaction with customers, and growing full-stack developers.
InfoQ: Can you elaborate about the benefits that you have seen in organizations that are using microservices?
George: The obvious benefit has been time to market, and the ability to capture corresponding windfall profits. And even bad ideas are detected quickly, minimizing the cost of going down the wrong path.
The surprising benefit has been the enablement of innovation. The business, after the initial "Wow!" of the rapid delivery, started coming up with new ideas. With the burden of long lead times removed, new ideas started flowing. "Could we try…" became common, with positive answers. The business was re-energized.
InfoQ: What do you expect that the future will bring in microservices?
George: I believe we will see rapid maturing of this architecture. Books and patterns will be published. Dedicated conferences and consultants will arise. And, unfortunately, there will be introductions of processes and procedures to "standardize" microservices to control those environments; the net of which, of course, will be a reduction in the twin benefits of rapid delivery and innovation.