sukun

My theory here is. Say we are measuring TCP+IP6(tcp6) success rate. - The peer that was found by honeypot had a relay address. So its reachability was private and it'd...

We have 1 fix in: https://github.com/pion/mdns/pull/172 We need 1. Remove the closed datachannels from the `pion/webrtc.PeerConnection` object. 2. Fix a memory leak in pion/sctp

Tracking issue for 1. https://github.com/pion/webrtc/issues/2672

For 1. Remove the closed datachannels from the pion/webrtc.PeerConnection object The PR in webrtc decreases memory usage from 31MB to 17MB when running 10 connections and echoing 1MB over 100...

Setup: two ec2 instances. c5.xlarge(4 cores 8GB Ram). us-east-1 and us-west-2. Ping time is 60ms. BDP is 40MB assuming 5Gbps link. - All bandwidth numbers are in Mega Bits per...

I quite like the idea of 100kb receive buf. The performance seems acceptable for now and the peer can enqueue 10x less on the sctp layer.

@SgtPooki the better way to do that is to limit the sctp receive buffer to 100kB as this affects all streams on a connection. Apologies for the poor naming, I...

That's a much larger change since the underlying sctp connection's receive buffer is a fix sized buffer and doesn't support window updates.

Sure. I can take this. I think the last time I looked at this my understanding was that this is normal IPFS traffic since the packet sizes mentioned in the...

For source address spoofing, I'm happy with the provided config options. A config option that's number of concurrent connections per IP might be nice for added DoS protection. But that's...