I created a tcpdump file:
tcpdump -i eth0 host xxx.208.xxx.59 -n -s 0 -vvv -w /tmp/dump.dmp
duration was about 3 hours.
This file now has 450 MB. Can I say now that the IP xxx.208.xxx.59 generated 450 MB traffic in 3 hours?
I created a tcpdump file:
tcpdump -i eth0 host xxx.208.xxx.59 -n -s 0 -vvv -w /tmp/dump.dmp
duration was about 3 hours.
This file now has 450 MB. Can I say now that the IP xxx.208.xxx.59 generated 450 MB traffic in 3 hours?
Yes, maybe, not necessarily.
A pcap file is not simply a byte-for-byte representation of the traffic that was sent/received. Things that will contribute to inaccuracies include:
If you want to account for traffic, do it properly, with port or netflow statistics retrieved from your core.
I would say yes. As it is my understanding the writer (-w) writes the packets byte-for-byte to /tmp/dump.dmp. But I'm only 80% sure...
That would include header information also, but that should be calculated into the throughput statistic.