After a little discussion on my dev discord about the overhead of JSON REST communication between microservices, I was curious how much overhead pure HTTP JSON REST communication actually introduces.
The test setup consists of two services, one as server and one as sender/receiver. Both inside Docker containers communicating over a private Docker network.
The sender crafts a JSON REST request, sent to the server which deserializes the received payload, adds timestamps and then serializes the object again to a JSON payload responded to the sender.
The sender then repeats this and measures the average RTT inclusive the serialization and deserializition times of the sender itself.
The following results below were tested on Docker for Windows (WSL2) v20.10.6.
