Reconsider payload wrappers #555

New issue

Open

opened 2024-11-18 12:43:51 +00:00 by alexvanin · 0 comments

alexvanin commented

2024-11-18 12:43:51 +00:00

Owner

This issue requires some research and benchmarks at first.

We have plenty payload wrappers that:

calculate hash sums
determine object type based on payload,
control the size of the sending data chunk in sdk-go, etc.

Latest payload wrapper is introduced in #551

While having many wrappers simplifies the job for programmer by providing simple abstraction over payload data, I think it may cost a bit of performance during heavy I/O. We need to benchmark it to investigate.

The biggest issue for me is that all those wrappers have different buffer size. Some buffer size alignment may help to improve performance. Together with it, remove all hard-coded buffer sizes for wrappers. Reuse existing config parameters or define a new one.

This issue requires some research and benchmarks at first. ## Is your feature request related to a problem? Please describe. We have plenty payload wrappers that: * calculate hash sums * determine object type based on payload, * control the size of the sending data chunk in sdk-go, etc. Latest payload wrapper is introduced in #551 While having many wrappers simplifies the job for programmer by providing simple abstraction over payload data, I think it may cost a bit of performance during heavy I/O. We need to benchmark it to investigate. The biggest issue for me is that all those wrappers have different buffer size. Some buffer size alignment may help to improve performance. Together with it, remove all hard-coded buffer sizes for wrappers. Reuse existing config parameters or define a new one.