I learnt something new today courtesy of the CEO of a partner I met today. Over the last two weeks the subject if micro-services has crept into the discussions and emails. So when the CEO turns out to be an expert on micro-services I asked him to explain.
In his words a micro-service is “a service that can be built in a week of less”. This approach of quick time to market fits very well with the philosophy behind the IBM platform as a service (PaaS) IBM Bluemix. That is, to be able to create and deploy applications quickly, quick time to failure through iterative development and testing and ultimately quick time to market.
The new world of code development for the cloud is an area I am increasingly learning about as I work with partners who want to use IBM Bluemix and evolve content for the IBM Markeplace.
Following on from the idea that micro-services are quick to build is the fact that they are also easierto maintain, debug and support. With many micro-services potentially making up a broader application or service, you now have a set of discrete services that overall enables easier debugging maintenance and support. These many modules are smaller in size and less complex than larger modules therefore, in theory, if one fails it is less likely to destabilise the overall system. Or if it does then it is quick to fix or replace. This concept is know as “anti-fragility”.
Now the notion that something is potentially replaceable is what makes mcro-services and “anti-fragility” all the more interesting as only those micro-services that you know to be reliable and trusted start to be the winners in the market.
Let the market place determine the outcome.