Explain interfaces with over 100 percent utilizations

[lchance]

I use and review various NMS platforms for WAN/LAN monitoring and management and often get asked why an interface(s) such as Load Balanced T1's or ATM or other interfaces report utilizations over 100%...such as, 120% or 180%, etc.

Anyone have any ideas or references to call on that explain these by formulas or such? I've studied these as best I can but never get any solid support to match my answers or theories, especially when influences include cache, duplex and or when ingress and egress traffic patterns are involved.

Thanks.

[Richard Burts · Accepted solution]

Larry

There may be several answers to your question depending on the context in which it is asked. As you suggest one possibility is involving duplex. If you have a T1 that has 1 Mb outbound and has 1 Mb inbound are you calculating 2 Mb of load and characterizing it as over 100 %? Obviously the answer there is that with media that are full duplex the real capacity is 2 x stated capacity.

Another possibility is associated with certain media that may be burstable such as Frame Relay. You may have contracted for a 256 Kb PVC. But you are able to send more than 256 Kb of data since Frame Relay does generally have a committed rate and a burstable rate.

Another possibility is that the interface may have been configured with a bandwidth value that does not really reflect the capacity of the interface. The NMS reports are generally using the configured bandwidth to determine the capacity and if the real capacity is greater than the configured bandwidth then the reports will show greater than 100 % utilization.

HTH

Rick