On Fri, Oct 3, 2008 at 8:49 AM, Peter Van Epp <[EMAIL PROTECTED]> wrote: > Doesn't appear to be a network issue. All packets presented appear > to have been sent without delay of complaints about lack of buffer space. > That points at disk (which is, assuming a properly configured network) usually > the bottleneck anyway. A couple of suggestions: use iperf/netperf between > the hosts to verify the network will go fast for some value of fast (the > next most likely bottleneck is DDR2 ram at 6.4 gigs per second and/or the > PCI bridge chips on the motherboard) at least faster than 380 megabits per > second (we can achieve 995 megabits per second across 3000 mile gigabit paths > with properly tuned linux boxes so this shouldn't be an issue). Then you are > in to disk I/O benchmark software such as bonnie to see what your disk > subsystem read speed is (I'd guess around 380 megabits per second :-)). This > is limited by spindle seek time primarily and a few other factors such as > thermal recalibration. Performance is a serial process, it is only as fast > as the slowest point in the path.
Since he's using the memory cache and a pretty sizeable loop count, it's probably not disk I/O. I'm guessing most of it is overhead of doing 500K write()'s per second. When you're getting 995Mbps, I doubt your average packet size is 102 bytes like Jeff's. If it is, then I'd love to hear more about your setup (kernel version, tuning, hardware, etc) -- Aaron Turner http://synfin.net/ http://tcpreplay.synfin.net/ - Pcap editing and replay tools for Unix & Windows They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety. -- Benjamin Franklin ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Tcpreplay-users mailing list Tcpreplay-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tcpreplay-users Support Information: http://tcpreplay.synfin.net/trac/wiki/Support
