2009/12/4 Les Mikesell <[email protected]>: > I'm too lazy to do the math, but those graphs look fairly predictable just > from the number of packets and retry times involved - at least considering > the syn retransmit timing spec and tcp retry backoff.
It's been a while since that presentation, and I don't remember the precise details. But from what I do remember, the main results from Wes' experiments revolved around the step-changes in behaviour as the error rate increased. So yes - the graphs do look fairly predictable at first sight. Lots of roughly parallel lines, with generally similar behaviour. But one of Wes' discoveries was that at a certain level of traffic corruption, particular configuration of request characteristics (either TCP-based queries or UDP timeout/repeat parameters) suddenly gave dramatically worse behaviour. But also that for mild-to-medium lossy networks, the behaviour of SNMP-over-TCP was nowhere near as bad as received wisdom had tended to assume. But you'd really need to ask Wes himself, to get the full details. Dave ------------------------------------------------------------------------------ Join us December 9, 2009 for the Red Hat Virtual Experience, a free event focused on virtualization and cloud computing. Attend in-depth sessions from your desk. Your couch. Anywhere. http://p.sf.net/sfu/redhat-sfdev2dev _______________________________________________ Net-snmp-users mailing list [email protected] Please see the following page to unsubscribe or change other options: https://lists.sourceforge.net/lists/listinfo/net-snmp-users
