Performance

xsrpcj has been written with performance in mind, it uses a simple and efficient communication model natively on top of sockets.

As part of the example applications, a performance comparison test has been done with Grpc java, the native RPC implementation for protocol buffers.

The performance test is a simple application scenario doing many rpc calls and measuring the required time. It is an indicative performance test.

You can run the tests on your reference machine, in principle the results are as follows:

• for small data interactions xsrpcj performs significantly faster. Many small data interactions bring out the processing delays of the framework itself & the network transport packet size vs payload size. In this scenario xsrpcj performs very well in comparison.
• for large data interactions xsrpcj performs marginally faster. Here the framework delays & the payload vs overall packet size are dwarfed by the network delays, which is the dominant factor in large data interactions.

xsrpcj tries to be an extra small and ultra fast RPC implementation