Add option to measure statistics which are dumped or returned on query.

[fruffy]

An example of useful statistics can be found here:
https://github.com/flowgrind/flowgrind

[fruffy]

Ideally this will be implemented by capturing packets and providing a summary of the capture over the last n minutes. This saves disk space and shows a snapshot performance of all algorithms.

[tomyang729]

todos:

1. using hostname rather than ip addr in the logging
2. convert the logging file to csv files
3. measure rtt for each connection. So we need to have a separate icmp connection per data connection??