Burkhard Kaas writes: > we want to monitor packet transit time and packet loss rate. Packet > round trip time (RTT) and ping loss rate are not sufficient for us, > because we don't see which direction causes variations in RTT and > packet losses.
> Does anyone know about tools do do this? I know about the problems > with accurate time synchronization, we have GPS time sync with > accuracy better than 100us available. It would be perfect to view > the results with smokeping. I know several systems that measure one-way delay and loss. As you say, one-way delay requires synchronisation (one-way loss is easy though). Both require special software at both ends. rude/crude - http://rude.sourceforge.net/ Simple command-line tools to send and receive UDP packets. The sender generates packets with timestamps, according to either a rate/size, or a script with packet sizes and relative timestamps. The receiver generates a log file including sender timestamps (from the packets) and receiver timestamps, which you could then use to feed something like SmokePing. It shouldn't be too difficult to build a system based on this yourself. OWAMP - http://e2epi.internet2.edu/owamp/ Something like rude/crude, but running as a daemon with some signaling mechanism to initiate measurements. I don't personally have any experience with this. RIPE TTM - http://www.ripe.net/test-traffic/ An entire network of systems running one-way delay and loss measurements (and traceroutes). You can order one or several boxes and have RIPE NCC run them for a fee. Measurement data is processed centrally in Amsterdam, with useful analyses, nice graphical representations, and good query (and on-demand graphing) options. Normally all measurements from all boxes is publicly accessible. I'm sure there are others. Hope this helps, -- Simon. -- Unsubscribe mailto:[EMAIL PROTECTED] Help mailto:[EMAIL PROTECTED] Archive http://www.ee.ethz.ch/~slist/smokeping-users WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
