Cisco Cisco IOS Software Release 12.2(33)SRE

Descargar
Página de 58
RSVP Fast Local Repair
  Information About RSVP FLR
3
Cisco IOS Release 12.2(33)SRB
Figure 1
Overview of RSVP FLR
Initial RSVP states are installed for an IPv4 unicast flow over Routers A, B, C, D, and E. Router A is 
the source or headend, while Router E is the destination or tailend. The data packets are destined to an 
address of Router E. Assume that a route change occurs, and the new path taken by the data packets is 
from Router A to Router B to Router F to Router D to Router E; therefore, the old and new paths differ 
on the segments between Routers B and D. The Router B to Router C to Router D segment is the old 
segment, while the Router B to Router F to Router D segment is the new segment. 
A route may change because of a link or node failure, or if a better path becomes available. 
RSVP at Router B detects that the route change affects the RSVP flow and initiates the FLR procedure. 
The node that initiates an FLR repair procedure, Router B in 
, is the point of local repair (PLR). 
The node where the new and old segments meet, Router D in 
, is the merge point (MP). The 
interfaces at the PLR and the MP that are part of the old segment are the old interfaces, while the 
interfaces that are part of the new segment are the new interfaces.
If a route has changed because of a failure, the PLR may not be the node that detects the failure. For 
example, it is possible that the link from Router C to Router D fails, and although Router C detects the 
failure, the route change at Router B is the trigger for the FLR procedure. Router C, in this case, is also 
referred to as the node that detects the failure.
Benefits of RSVP FLR
Faster Response Time to Routing Changes
FLR reduces the time that it takes for RSVP to determine that a physical link has gone down and that the 
data packets have been rerouted. Without FLR, RSVP may not recognize the link failure for 30 seconds 
when all of the sessions are impacted by having too much traffic for the available bandwidth. With FLR, 
this time can be significantly reduced to a few seconds.
After detecting the failure, RSVP recomputes the admission control across the new link. If the rerouted 
traffic fits on the new link, RSVP reserves the bandwidth and guarantees the QoS of the new traffic.
If admission control fails on the new route, RSVP does not explictly tear down the flow, but instead sends 
a RESVERROR message towards the receiver. If a proxy receiver is running, then RSVP sends a 
PATHERROR message towards the headend, in response to the RESVERROR message, indicating the 
admission failure. In both cases, with and without a proxy receiver, the application tears down the failed 
session either at the headend or at the final destination.
Until this happens, the data packets belonging to this session still flow over the rerouted segment 
although admission has failed and QoS is affected.
Old interface
Old segment
New interface
PLR
PLR = point of local repair
MP = merge point
MP
170682
Old interface
New interface
A
B
C
D
E
New segment
New segment
F