I am keen on understanding the following aspects of micro-services.
Commit and roll-back. If a business goal is to be achieved by invoking multiple micro-services, who co-ordinates commit and roll-back? Are micro-services required to embed the commit-roll-back protocol in their implementation? Or, is it provided by some external factor?
Load-balancing for micro-services. Is a micro-service front-ended by a load balancer to help it scale? And, therefore, scaling is out of scope of implementation of the service?
Co-relation with AWS Lambda/Bluemix OpenWhisk/Azure Functions. How do micro-services compare with ‘serverless’ architecture?