Serialization format of gdb trace data between many different machines.

[joshbaptiste]

I have a corewatch.py program on ~400+ Linux machines that collect
core file gdb output and sends the output traces via email to
developers. I what to convert the script for all output to be sent to
a centralized server also written in python and displayed via
webpage.  My first thought was a simple HTTP POST to server from all
nodes but since the gdb output can be very large, multiplied my the #
of servers, worries me as network traffic of 400+ nodes sending to 1
server may cause saturation.

What would be the most efficient/small serialization format to use to
send this trace data from nodes to server over network.. would
Pickling be ok?