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
There is a chance to try UDT on real world events:
Source code is simple (good+), and udt is a challenging protocol that is/was leading, but still untested or unkown by the masses:
let's make some tools out of this git and clarify for life:
is udt still a challeging and good protocol, competitive to tcp-ip, or mux ?
or it's today simply outdated;
well we can sort it out:
Make two tools:
sendclient.c and receiveclient.c -> refactored and updated to match IPERF
results: achieve on localhost speeds up to 10 and 99GBPS as iperf does on a laptop or commodity hardware;
sendfile.c and receivefile.c -> make a connection between two vps, from japan to us and achieve better results than a tunnel, less cpu% consumption and much faster than kcp-tun (go) and/or tsunami
As for myself, I can publish benchmarks for the current UNPATCHED and unmodified versions of UDT quite fast, can we do that ?
So, is making UDT a world-class protocol still feasible ? Does it pay the effort ?
Can we publish results as they are from git ? Or: is there another project from the same (or near-like) authors that surpasses this git project ?
Thanks all for answering !
The text was updated successfully, but these errors were encountered:
There is a chance to try UDT on real world events:
Source code is simple (good+), and udt is a challenging protocol that is/was leading, but still untested or unkown by the masses:
let's make some tools out of this git and clarify for life:
or it's today simply outdated;
Make two tools:
sendclient.c and receiveclient.c -> refactored and updated to match IPERF
results: achieve on localhost speeds up to 10 and 99GBPS as iperf does on a laptop or commodity hardware;
sendfile.c and receivefile.c -> make a connection between two vps, from japan to us and achieve better results than a tunnel, less cpu% consumption and much faster than kcp-tun (go) and/or tsunami
As for myself, I can publish benchmarks for the current UNPATCHED and unmodified versions of UDT quite fast, can we do that ?
So, is making UDT a world-class protocol still feasible ? Does it pay the effort ?
Can we publish results as they are from git ? Or: is there another project from the same (or near-like) authors that surpasses this git project ?
Thanks all for answering !
The text was updated successfully, but these errors were encountered: