Network traffic

Hi, I’ve opened the IPFS interface and at the end of the section Status there is network traffic with the outgoing and incoming entries. I would like to know what exactly these entries indicate.

Thank you

Have a look at the documentation for ipfs stats bw:

By default it should show all traffic for all protocols and all peers. In is receiving and out is sending seen from your IPFS client.

Best regards


I would like to know if incoming means the files that other peers are storing on me and outgoing means the blocks that I am sending to those who need them.

Best Regard


1 Like

Sadly IPFS isn’t that far in development yet, that it will intelligently store any data on your computer at all.

The traffic that you’re seeing is just ‘beeing part of the network’ in terms of the connection is established and you’re part of index which is hold of all data stored on other computers.

If you want to serve certain data to other users, you can with taking part in a so called ‘Collaborative Cluster’.

A small tool called ipfs-cluster-follow will make sure, your IPFS always got the latest data of the cluster, and you’ll serve the data of the cluster to others who use the network.

Since this is a pretty new feature, there are not many options to choose from, but I’m sure in the short term there will be more available:

1 Like

Ok, thank you.
I have another question. Are there any applications that use Ipfs? If yes, could you tell me one?


1 Like

This is by design. IPFS will never send you files to store unless:

  1. You requested it
  2. You’re part of a cluster

This is by design. Otherwise, you can end up using space for nothing (maybe the file will never be requested), and you can end up storing data you don’t want to store. There would be legal issues otherwise.

IPFS clusters, on the other hand, will store data intelligently on your computer (especially collaborative clusters). But clusters are on an opt-in basis.

Yeah sure, but (judging from the version number) IPFS is early in it’s development. So this might change in the future. :slight_smile:

It might, but I don’t think it will much. Having the files cached by those who consume it enables having the caching to scale naturaly with the demand. Over-caching it would be wasteful, and ressourses won’t be allocated to store something useful (another content requested a lot). The exceptions would be archives (cluster should be fine with that), or content that needs to be close anytime, even if rarely asked for (CDN), but in that case you should pay for the extra burden (pay a pinning service or setup your own network of nodes). No need to use my normal user bandwith for that. Ot should remain opt-in.

Yes you’re right, but I would love to see IPFS adapt a Web of Trust, like adding people you know by public keys, connecting to them and give them the ability to store some data on your nodes.