[draft] experiments of using proto and add microbenchmarks

[Alva8756]

This is an experiments of using proto as GetComponents response payload.

The microbenchmark creates a simple response with 5 components. It creates a local client sending request against local server.

Micro-benchmark result:

BenchmarkGetComponentsJson
p50 latency: 3.412042ms
p90 latency: 4.685958ms
p99 latency: 19.750375ms

BenchmarkGetComponentsProto
p50 latency: 3.226333ms
p90 latency: 4.581083ms
p99 latency: 18.503042ms