I'm looking for an utility that could show me how much of a bandwidth each process are using. It should just print the data and exit as I want to feed that list into conky.
You can try NetHogs.
NetHogs is a small 'net top' tool. Instead of breaking the traffic down per protocol or per subnet, like most tools do, it groups bandwidth by process. NetHogs does not rely on a special kernel module to be loaded. If there's suddenly a lot of network traffic, you can fire up NetHogs and immediately see which PID is causing this. This makes it easy to indentify programs that have gone wild and are suddenly taking up your bandwidth.
atop ... to get the most out of it you may have to enable some additional kernel patches (I/O accounting patches).
atop isn't an option then use
netstat -anp --inet (as root) to provide a listing of which TCP/UDP ports are in use by which processes (or possibly use
lsof for that). From there simply iterate over each process that's got an open socket and either attach to it using
ltrace -S or
strace to look at the reads, writes, sends and receives, or use
tcpdump with a filter specifying your local IP address(es) and the TCP/UDP ports that were listed.
atop is certainly the most convenient of these ... if you have it and the necessary kernel support installed. In the past I've had customers and employers set up special systems (different from their production images) for the sole purpose of supporting I/O profiling using
atop. However these other techniques will get you there.
I'm sure we could also do something using SystemTap ... but I don't know of any easy pre-cooked recipes for doing this. SystemTap is very much a programming analysis tool.