Is there a way to setup variable QoS for a traffic flow? I have an 85 Mbps metro-ethernet WAN connection between two sites. One side has a Cisco 3745 router, the other has a Cisco 2821 router. We just recently installed two SANs (one at each site) and they are mirrored. I would like to be able to setup a max bandwidth usage of 50Mbps for the replication of the SAN. And then bump it back down to 25 Mbps while backups are taking place. Right now I'm using the "traffic-shape" command to limit traffic from the SAN to 25 Mbps by matching its IPs. If I left it at 50 Mbps, then that leaves about 30 Mbps for backups which slows down my backups.
In the future, I can probably get away with just having the 25 Mbps limit. But for now, I have to replicate several terabytes of data, so the more bandwidth I can give it the better.
The SAN uses iSCSI to replicate data, so I just match those IPs for limiting the bandwidth. The variable is really the time of day. So during the day, I run at 50 Mbps for the SAN, then at night it drops to 25 Mbps while backup are running. The reason I was looking for capping the SAN bandwidth was that it will consume the entire link if it's not throttled back. When that happens, response time for users drops considerably.
I had originally thought of something along these lines, but not sure if it will work.
Thanks. Does the percent value matter? Or is it the just ratio of the numbers?
I had looked into the time-based ACLs too. Would something like this work? The ACLs 101 & 102 would have the time-ranges set match to their class-map names. If the time is outside of the time-range set for an ACL, does it just ignore the ACL?
The percentage matters as it sets ratios to other traffic. For example
bandwidth percent 30
bandwidth percent 60
bandwidth percent 3
bandwidth percent 6
bandwidth percent 1
bandwidth percent 2
All the above will set the same ratios relative to each other, but what's the ratio relative to class-default? (BTW, many platforms/IOSs won't accept bandwidth percents that sum more than 75% unless reservation changed.)
Yes, the time-based ACL tied to classes should work. Of course, the policy-map doesn't really realize you're only changing bandwidth allocations for time of day. I.e. you've allocated 90% when you intend (and will effectively obtain with time-based ACLs) 30% or 60%. This can become annoying when you want to explicitly allocate bandwidth to other defined classes.
If your available downstream bandwidth is less than the interface's, don't forget a parent policy shaper. Otherwise, you'll usually end up (FIFO) queuing at the next bottleneck without the bandwidth allocations you intended. (This could explain poor performance you've seen for other traffic when SAN not capped with shaper.)
We are pleased to announce availability of Beta software for 16.6.3.
16.6.3 will be the second rebuild on the 16.6 release train targeted
towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are
looking for early feedback from customers befor...
Introduction Featured Speakers Luis Espejel is the Telecommunications
Manager of IENova, an Oil & Gas company. Currently he works with Cisco
IOS® and Cisco IOS XE platforms, and NX to some extent. He has also
worked as a Senior Engineer with the Routing P...
In this session you can learn more about Layer 3 multicast and the best
practices to identify possible threats and take security measures. It
provides an overview of basic multicast, the best security practices for
use of this technology, and recommendati...