Hi Steve, 

Perhaps the example I used of comparing microservices to breaking up a framework is too simplistic and a bit contrived.  There is no direct comparison.  You're not going to gain anything by breaking up a framework into its basic parts. That said, what do you do when you have a large company that needs to share data across multiple services. It is also a bit simplistic to assume that you're just going to use WordPress for one service and Drupal for another and then tie them together in some way (which I have done) so that you can minimize the code you write and stay with the OOP design of each.  In both cases most likely you will be using about 1 percent of the capability of each framework but because of the way both products' event looping works you will be wasting 90 percent of each server's resources to complete each request, costing the company thousands if not hundreds of thousands of dollars per month in extra compute time usage (the real gross part).

So software engineers build small single purpose services with lightweight libraries allowing the interpreter to parse a thousand lines of code rather than a million while making 1 tenth the number of requests - thereby operating on tiny containerized instances of Linux.  And then someone packages these lightweight services  as a SaaS based control plane product and then sells compute services online and now anyone can connect services together that can both rival the most used WordPresses  plugin system but with a service that can scale to 2 to 3 orders of magnitude more any traditional CMS framework model, without spinning up any new services.  

Netflix is not just one website framework but hundreds of dashboards (Views) that consume thousands of endpoints each controlled by a microservice.  It is estimated that one third of the internet usage during any given week night after 7pm is requests to Netflix streaming services - which are all handled by thousands of microservices.