You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I know youre looking at metrics (ipfs/kubo#1418) but I just figured i'd file and issue here to discuss what numbers we want to see at the end of a bitswap test.
I have a couple numbers built into the 'stat' command, such as total blocks received and the number of duplicate blocks received (duplicate blocks is arguably the most important metric right now). But also important is mean time from block request to block receipt, and overall bandwidth across varying network conditions. Also nice to know would be bandwidth overhead, how much extra data are we sending on top of the actual file being requested.
The text was updated successfully, but these errors were encountered:
Would you like mean block time per bitswap instance, mean block time across all bitswap instances, or both?
And could you give me a pointer on how I would get bandwidth information? Would it make sense to change the ReceiveMessage function in bitswap to log the received message info somewhere? Or could I have my own stream reader that logs the number of bytes it reads in mocknet?
Would you like mean block time per bitswap instance, mean block time across all bitswap instances, or both?
both
And could you give me a pointer on how I would get bandwidth information? Would it make sense to change the ReceiveMessage function in bitswap to log the received message info somewhere? Or could I have my own stream reader that logs the number of bytes it reads in mocknet?
bandwidth can come from the connections. we already have metrics like these, i think. @whyrusleeping ?
I know youre looking at metrics (ipfs/kubo#1418) but I just figured i'd file and issue here to discuss what numbers we want to see at the end of a bitswap test.
I have a couple numbers built into the 'stat' command, such as total blocks received and the number of duplicate blocks received (duplicate blocks is arguably the most important metric right now). But also important is mean time from block request to block receipt, and overall bandwidth across varying network conditions. Also nice to know would be bandwidth overhead, how much extra data are we sending on top of the actual file being requested.
The text was updated successfully, but these errors were encountered: