Hi all
We have quite a large number of branch sites in a hub-and-spoke DMVPN topology via mpls links. Recently we have experienced significant slowness to many of these sites along with packet drops on our hub interface. We are doing qos via our NHRP groups and are aware that there are quite a few limitations with this. Our branch sites have fairly low bandwidth links so we are frequently having to mess around with these in order to improve performance for one vlan at a branch site over another.
We've been recommended to increase our queue-limit on our hub from the default of 64 packets to 128 or 256, however I've tried doing this and it has informed me that this cannot be done on a multipoint tunnel interface. We are also limited to only using shape-average. Is there any other way we can tweak our policies/queues to reduce these packet drops we are experiencing? The obvious solution would be increasing our bandwidth at our tails however this is not really an option right now.
No comments:
Post a Comment