IPFS is a resource hog. Not sure if it's just the implementation or if it's because of architectural limitations, but it gobbles up CPU and RAM much more than you'd expect it to. I shut down my self-hosted IPFS server because it kept pegging a CPU core doing what seemed like nothing (no new pins, barely any network traffic, just sitting there).

(Quiet founder here) Yes, IPFS is a resource hog when running on the global IPFS network, but Quiet never touches that network, and this improves performance a lot.

I can run Quiet in the background all day on my Pixel 6 and we haven't even started optimizing for battery life yet.

Client side compute for other random people is probably what GP is talking about here.

IPFS doesn't do that, are you thinking of something else?

Did you try to look at IPFS codebase size?

  ~/kubo$ find . -name \*.go | xargs wc -l | tail -1
  59629 total

Doesn't seem that big?

I'd like to point out that you probably cloned https://github.com/ipfs/kubo there, which is only one part of what accounts for the "IPFS codebase". If you want to find out the actual LoC, I'm afraid you will need to dig through https://github.com/ipfs/kubo/blob/master/go.mod and pick everything that's github.com/ipfs/*, github.com/libp2p/* and github.com/ipld/*.

Fair point!

YES! Compared to a simple IRC server+client or XMPP definitively.

Well irssi is an IRC client and is bigger than that:

  ~/irssi$ find . -name \*.[ch] | xargs wc -l | tail -1
  88551 total

https://github.com/irssi/irssi

Obviously ipfs doesn't have the chat UI, ipfs is just a protocol layer.

But it is a more complicated protocol than IRC, so you'd expect it to be larger. I don't think it's large out of proportion, certainly not large enough to call it a "MONSTER software".

My users complain all the time about the LoC of my codebase