Hello Netplace Support,
the load interval provides a time interval used for calculation of
packets /s
bytes /s
both input and output direction.
These counters do not provide the instant value they are rather an exponentially weighted average.
The default load interval values of 5 minutes requires 20 minutes (4 times the timer) for the counters to reflect the effective bit rate when sending a costant bit rate in a lab test.
If in the lab test I send 20 Mbps of traffic the input interface counter will tell us that input bit rate is 20 Mbps after 20 minutes (4 times the default load interval).
When using load interval 30 seconds the convergence time is reduced to 2 minutes so in a lab test the input rate will become 20 Mbps after two minutes ( 4 times the timer).
The great benefit is that when we issue a show interface with load interval 30 the counters are more able to follow the current traffic speed.
>> What is the use of command load-interal 30 on Cisco router interface & how this is helping to load share along with Bandwidth cmd
I have explained what the load-interval command does. The bandwidth command is actually and administrative command that can reflect the current speed of the interface or it is set to a value with a configuration command.
The bandwitdh parameter is used by EIGRP in EIGRP composite metric calculation.
the bandwidth parameter is used also as a reference for the bandwidth by modular QoS framework when configuring CBWFQ with or without priority queue ( LLQ).
EIGRP in theory should be able to use load levels on interfaces as an optional parameter in composite metric calculation.
In reality when it comes to load and reliability, these features have never been implemented completely.
If the K values for load and /or reliability are enabled in EIGRP the values are taken as a static snapshot at the time that EIGRP is enabled on the interface. To avoid oscillations and routing instability these values are not updated dynamically to reflect the current load or reliability. Even if this is declared as an EIGRP capability in marketing documents.
When talking about PfR performance routing, in this case the load level is tracked and having load interval set to 30 seconds provide the benefit of more reactive / less averaged values.
Hope to help
Giuseppe
