Hi Alejandro, If you can run 15.0 code you may want try HQF, rather then CBWFQ.
There is a good description at the link below: http://blog.ioshints.info/2009/11/first-hqf-impressions-excellent-job.html?utm_source=feedburner&utm_medium=RSS&utm_campaign=IOS+hints+Feed Hope it helps, Dave On Fri, Nov 20, 2009 at 5:51 PM, Alejandro Selios <asel...@gmail.com> wrote: > Thanks Clue for your quick answer. > > In the original mail I didn't mention it, but I applied the service-policy > to the PVC like you said. (Below attach the complete configuration for both > cases). > > Another issue is that Cisco's 800 and 1800 series routers only have 2 level > of prioritization. Cisco's routers allow to configure a priority queue and > CBWFQ (class base weighted fair queue) but does not has the flexibly to > specify a scheduler policy or any other queuing policies. > > I think that weighted fair queue will leave as dead end, I think the way to > follow is a strict priority queuing with a customizable scheduler policy. > I'm looking for a way to do that in Cisco's routers. > > I'll appreciate any help you can give me. > > > Cisco 878/877 > -------------------------------------- > class-map match-all GOLD > match access-group name GOLD_ACL > class-map match-all SILVER > match access-group name SILVER_ACL > > > policy-map QoS_POLICY > class GOLD > priority percent 25 > police cir percent 25 pir percent 25 > set dscp ef > class SILVER > set dscp af41 > bandwidth percent 50 > shape average percent 75 > class class-default > fair-queue > set dscp af12 > bandwidth percent 25 > > interface ATM0 > no ip address > load-interval 30 > no atm ilmi-keepalive > max-reserved-bandwidth 100 > ! > interface ATM0.35 point-to-point > description INTERFAZ DE ACCESO > ip address x.x.x.x 255.255.255.254 > atm route-bridged ip > pvc 0/35 > vbr-nrt 576 576 > oam-pvc 0 > encapsulation aal5snap > service-policy output QoS_POLICY > > > > Cisco 1841 > -------------------------------------- > > class-map match-all GOLD > match access-group name GOLD_ACL > class-map match-all SILVER > match access-group name SILVER_ACL > > > policy-map QoS_POLICY > class GOLD > priority percent 25 > police cir percent 25 pir percent 25 > set dscp ef > class SILVER > set dscp af41 > bandwidth percent 50 > shape average percent 75 > class class-default > fair-queue > set dscp af12 > bandwidth percent 25 > > policy-map SHAPER > class class-default > shape average 1000000 > service-policy QoS_POLICY > ! > ! > ! > interface FastEthernet0/0 > description INTERFAZ DE ACCESO > no ip address > load-interval 30 > duplex auto > speed auto > ! > interface FastEthernet0/0.100 > encapsulation dot1Q 100 > ip address x.x.x.x 255.255.255.252 > service-policy output SHAPER > > > > Thanks, > > > > > > > 2009/11/20 Clue Store <cluest...@gmail.com> > >> Hi Alejandro, >> >> If you're doing this with xDSL, you will have to apply the service-policy >> to the PVC for outgoing to make it work. Here's an example of what I use. I >> know your class-maps and policy-maps are different, but you should be able >> to get the point. >> >> class-map match-any VoIP >> match ip rtp 16384 16383 >> match access-group name VoicePorts >> >> policy-map Voice >> class VoIP >> priority percent 75 >> >> class class-default >> fair-queue >> >> >> interface ATM0.1 point-to-point >> pvc 1/100 >> vbr-rt 384 384 >> tx-ring-limit 3 >> service-policy output Voice >> >> >> ip access-list extended VoicePorts >> permit udp any host x.x.x.x range 22026 62025 >> permit udp any host x.x.x.x range 22026 62025 >> >> >> >> >> HTH, >> Clue >> >> On Fri, Nov 20, 2009 at 9:26 AM, Alejandro Selios <asel...@gmail.com>wrote: >> >>> Hi, >>> >>> I want to configure QoS on a Cisco's 800 and 1800 series router and I >>> couldn't make it work the way I want. I've configured three classes of >>> service called GOLD, SILVER and BRONZE, designed for low latency, low loss >>> and data respectively. >>> >>> * I want the GOLD class to have a maximum bandwidth of 25%, even if there >>> is >>> traffic in the other classes. >>> >>> * The SILVER class has a minimum bandwidth of 50% and, if there's no >>> traffic >>> in the GOLD queue, I want this class to take the available bandwidth which >>> is not used by the GOLD class. Thats means that the SILVER class has a 50% >>> of assured bandwidth and can increase to a maximum of 75% of the total >>> bandwidth. >>> >>> * The BRONZE class has no assured bandwidth and can only take the >>> bandwidth >>> which is not used for the other classes. >>> >>> The problem is that with Cisco's QoS tools I have not reached the desired >>> goal. I have tried with different configurations and IOS and I couldn't >>> get >>> any successful result . Below I attach the policy configuration that I >>> have >>> tried (just to give you an example). >>> >>> Is there any way in which I can achieve the behavior that was described >>> above?. >>> >>> Thanks in advance. >>> >>> >>> Standard Configuration >>> ------------------------------------- >>> >>> policy-map QoS_POLICY >>> class GOLD >>> priority percent 25 >>> police cir percent 25 pir percent 25 >>> set dscp ef >>> class SILVER >>> set dscp af41 >>> bandwidth percent 50 >>> shape average percent 75 >>> class class-default >>> fair-queue >>> set dscp af12 >>> bandwidth percent 25 >>> >>> >>> >>> >>> Hierarchical Configuration >>> --------------------------------------------- >>> policy-map GOLD+SILVER >>> class GOLD >>> priority percent 33 >>> police rate percent 33 >>> class class-default >>> fair-queue >>> >>> policy-map QOS >>> class GOLD+SILVER >>> shape average percent 75 >>> bandwidth percent 75 >>> service-policy GOLD+SILVER >>> class class-default >>> fair-queue >>> >>> >>> Note: This configuration was applied in a subinterface (xDSL, 802.1q >>> ethernet) and for the outgoing traffic. >>> _______________________________________________ >>> cisco-nsp mailing list cisco-...@puck.nether.net >>> https://puck.nether.net/mailman/listinfo/cisco-nsp >>> archive at http://puck.nether.net/pipermail/cisco-nsp/ >>> >> >> > _______________________________________________ > cisco-nsp mailing list cisco-...@puck.nether.net > https://puck.nether.net/mailman/listinfo/cisco-nsp > archive at http://puck.nether.net/pipermail/cisco-nsp/ > _______________________________________________ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/