Provide architectural and other technical guidance for TCP/IP RPC based programming
We have several legacy TCP/IP RPC, Client/Server programs that we must port to Azure Service Fabric; however, support and guidance when migrating TCP/IP program to SF Cluster are not supported. We are not told we can't do it, but when it gets down to actual development and testing support, it turns out that quite a bit of design work and coding is necessary to port an RPC program to SF. One of the primary issues is when your RPC server generates IP port assignments at runtime; this type of design does not work very well in SF as each listener can only listen on a single port. To get this to work now, you must properly configure that Azure LB, and you must refactor your RPC server-side code to integrate with SF naming service, port assignment, port collision, and a few other issues.