Manifetch PoC: Compare performance

[ajnavarro]

We want to test if a new protocol able to get CIDs from a Root CID can improve somehow BitSwap performance.
From the work done by Adin here: #5 we need to do the following steps:

• Fix problems with the manifetch implementation, sometimes it gets stuck when processing CIDs
• Get some results on a table to be able to compare them. Work in progress here: https://docs.google.com/spreadsheets/d/1jEI8MpG8bodKcmXp5aJMG-h1grG6FUo5NKdEO5WP-xk/edit#gid=0
• Create some graphs comparing different DAG types, bitswap with and without manifetch...

[BigLep]

@ajnavarro : for the results table:

1. It would be ideal to have everything be in the same unit (e.g., seconds) so that can do a sort and easily compare.
2. Is there code we can link to that shows the test setup. Ideally we're running each test 3+ times and taking the average.

[ajnavarro]

@BigLep I took note and modify the spreadsheet to reflect the changes. Thanks!

[BigLep]

@ajnavarro : I think we can push this to Kubo 2023 for now until this work gets prioritized. Agreed?

[ajnavarro]

@BigLep 👍