Hi, I'm writing a backend plugin to obtain a daily AS-AS traffic matric in my network, with 38 ASs and 21 sources. The only way I found so far is to get the information with nfdump (1.5) running #nfdump -M <source_list> -R nfcapd.$tart_tslot:nfcapd.$end_tslot -n 50 -s srcas/bytes -o long "src as $src_as and dst as $dst_as"
as many times as all the possible AS-AS pairs (38X38), and then parse the output. Note that I use -n 50 but I could vell have used -n 1 (because of the filtering, I always get that there is only one contributing src_as). The problem is that given the number of flows (roughly: 300 k flows per source and per hour, with each AS connected to one, or two, or three sources at most), the processing time is high. I probably won't be able to run the processing every day over all the past 24 hours, but I'll be forced to focus on a limited time slice. Questions: 1) is there another easy way to do? 2) if not, how difficult would it be (and what module should be modified) to let nfdump have prallel filters? The processing bottleneck is clearly the disk access bandwidth (the cpu stays at about 4-5%). Regards, Maurizio ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Nfsen-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/nfsen-discuss
