OSPF Downward Option

PE1 router advertises intra area routes int MP-BGP. other PE2 redistributes the MP-BGP routes into OSPF as LSA type 3 (Summary address) as MPLS backbone acts as an OSPF superbackbone.

PE2 sets the Down bit on those routes to avoid routing loops in scenarios which a site is dual homed to the service provider network.

If a router receives a Type 3 LSA with its Down bit set on an interface that belongs to a VRF, it drops the LSA. In cases which the site is not dual home this loop prevention mechanism is undesirable.

To disable this behaviour the following command can be used under router ospf, which no longer perform PE specific checks:

R1(config-router)# capability vrf-lite

If the router doesn't support that command, different OSPF domain-ids should be set on the PE routers, which in turn changes the route types to External(LSA Type 5) instead of Summary (LSA Type3).

max-reserved-bandwidth

To change the percent of interface bandwidth allocated for Resource Reservation Protocol (RSVP), class-based weighted fair queueing (CBWFQ), low latency queueing (LLQ), IP RTP Priority, Frame Relay IP RTP Priority, Frame Relay PVC Interface Priority Queueing (PIPQ), or hierarchical queueing framework (HQF), use the max-reserved bandwidth command in interface configuration mode.


http://www.cisco.com/en/US/docs/ios/qos/command/reference/qos_m1.html#wp1054626

Frame Relay RTP Priority

map-class frame-relay <NAME>
frame-relay ip rtp priority <LOW PORT NUMBER> <PORT RANGE> <BANDWIDTH>

Don't forget to configure frame-relay fragmentation (FRF.12) as RTP Priority doesn't work without frame-relay fragmentation.
map-class frame-relay <NAME>
frame-relay fragmentation <SIZE>

Frame Relay Traffic Shape

R1(config)# map-class <CLASS-NAME>
R1(config-map-class)#frame-relay cir <CIR>
R1(config-map-class)#frame-relay bc <CIR>
R1(config-map-class)#frame-relay be <CIR>
R1(config-if)# frame-relay traffic-shaping

Apply to PVC:
R1(config-if)# frame-relay interface-dlc <PVC>
R1(config-fr-dlci)#class <CLASS-NAME>

Apply to interface:
R1(config-if)# frame-relay class <CLASS-NAME>


CIR=Bc/Tc

Be=(AR-CIR)*Tc

AR= Available Rate, the actual physical interface speed.


Don't forget to enable traffic shaping on interface first

R1(config-if)# frame-relay traffic-shaping





http://www.ine.com/resources/01700368.htm

Serial Link Compression

PPP Predictor:
Less CPU, more memory, less efficient.
R1(config-if)# compress predictor

HDLC Stacker (LZ):
More CPU, less memory, more efficient.
R1(config-if)# compress stacker

Frame Relay FRF.9 Compression
R1(config-if)# frame-relay map ip <IP-ADDRESS> <PVC> broadcast IETF payload-compression FRF9 stac one-way-negotiation

Selective Packet Discard

It applies a form of random packet drop to interface FIFO input queue.
It's a hidden command from IOS parser.

ip spd enable

ip spd mode aggressive --> bad packets are dropped as soon as the minimum threshold for hold-queue
is reached

spd headroom <x> --> BGP, IGP, L2 Keepalives

spd extended-headroom <x> --> IGP, L2 Keepalives

ip spd queue max-threshold <x>

ip spd queue min-threshold <x>

show ip spd

http://www.cisco.com/en/US/products/hw/routers/ps167/products_tech_note09186a008012fb87.shtml

Don't Miss the Permit Sequence in route-map

While modifying BGP route attributes with route-map, always remember to add permit line at the end of the list, or you'll end up filtering those routes which haven't been matched by previous route-map entries.