<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7653.38">
<TITLE>RE: [mrtg] Gigabit Interfaces, SNMP v2, and MaxBytes</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>mrtg@ying.net wrote on Fri 14-Nov-08 19:01:<BR>
>When corroborating data on the actual interface, should the 5 minute<BR>
>input/output rates from a 'sh int' be fairly close to what is reported<BR>
>by MRTG?<BR>
<BR>
not necessarily. The cisco 5 minute values are exponentially decayed to give more weight to the last 30 seconds, while the mrtg values are straight-line averages that get interpolated based on the time stamp.<BR>
<BR>
The only way to corroborate would be to look at packets in/out (also in show int .. counters), wait 5 minutes, look at the counters again, and then divide the difference by 300 to see if they matched.<BR>
<BR>
--<BR>
Dan McDonald, CCIE # 2495, CISSP # 78281, CNX<BR>
Austin Energy<BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>