Simmi Singla wrote: > > Hi > reliability 255/255 means the link that this interface has is > 100% reliable, no problems on link.Ideally this value should be > like this otherwise some problem on link. > Calcualted on scale of 1-255 as percentage
Wouldn't that be nice if it really were a percentage? :-) How many CPU cycles could it use to calculate a percentage? How long could it take a Cisco software engineer to develop the routine to convert it to a percentage, so we could view statistics on a Cisco router in a standard way that is similar to other tools. ;-) It's a fraction, not a percentage. It's a fraction of 255, where 255 is the maximum. I suppose 255 is a nice number for programmers. How about considering the user, though? Since you're from Cisco, I have to give you a hard time about this. :-) In the case of reliability, the interface keeps a moving average of the fraction of frames that arrive with an error compared to frames that arrive without an error. The types of errors it's concerned with are CRC, runts (too short), giants, overrun, no buffer, etc. > txload 74/255(indicates transmit load on the link) > rxload 14/255(indicated recieve load on the link) > > These values in addition to other depends on calculating cost > depending upon routing protocol. IGRP and EIGRP can depend on reliability and load in their calculation of cost. They don't by default, but they can if the metric weights command is used. > > Like take an example > Igrp/eigrp can > use Bandwidth,delay,load,reliability & mtu etc to calculate > cost of path.(they use these interface cost metrics Each interface has a reliability value. IGRP and EIGRP also send reliability info for each route. Reliability is the lowest reported realiability. Whereas delay is cummulative, reliabitity and bandwidth are the worst case. As is the case with all these Cisco averages, it's technially an exponential average over 5 minutes. You'll see that confusing phrase in much of Cisco's documentation. I researched what it meant, one gloomy Oregon afternoon. The software calculates a sample every 5 seconds and provides a moving average over a 5-minute period. A moving average incorporates feedback from previous samples into the current result. The most recent sample is weighted slightly more than previous samples, but because there are so many more previous samples, they contribute heavily to the result also. The weighting decays exponentially for each previous sample, but in 5 minutes, there are 60 previous samples, so they are still significant. The goal is to avoid skewing the result too much if the 5-second sample happens to take place when someone is downloading a huge MP3 file, for example, or when nobody was doing anything for whatever reason. Clear as mud, eh? :-) The details of the exponential average calculation are buried in a Cisco document titled "Definition of 'bits/sec' from 'sh int.'" The URL for the document is: http://www.cisco.com/warp/public/66/3.html. _______________________________ Priscilla Oppenheimer www.troubleshootingnetworks.com www.priscilla.com > > > > > ritul wrote: > > > > Hi all > > > > I had executed the command sh int s0 on one of my routers,in > it > > one of the > > line says > > reliability 255/255, txload 74/255, rxload 14/255 > > > > Can anyone explain me what does this mean? > > > > Thanks > > Ritul > > > > > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=62270&t=62236 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
