MPLS and RSVP
7450 ESS MPLS Guide Page 47
Reservation Styles
LSPs can be signaled with explicit reservation styles. A reservation style is a set of control options
that specify a number of supported parameters. The style information is part of the LSP
configuration. SR OS supports two reservation styles:
• Fixed Filter (FF) — The Fixed Filter (FF) reservation style specifies an explicit list of
senders and a distinct reservation for each of them. Each sender has a dedicated
reservation that is not shared with other senders. Each sender is identified by an IP address
and a local identification number, the LSP ID. Because each sender has its own
reservation, a unique label and a separate LSP can be constructed for each sender-receiver
pair. For traditional RSVP applications, the FF reservation style is ideal for a video
distribution application in which each channel (or source) requires a separate pipe for each
of the individual video streams.
• Shared Explicit (SE) — The Shared Explicit (SE) reservation style creates a single
reservation over a link that is shared by an explicit list of senders. Because each sender is
explicitly listed in the RESV message, different labels can be assigned to different sender-
receiver pairs, thereby creating separate LSPs.
Note that if FRR option is enabled for the LSP and selects the facility FRR method at the
head-end node, only the SE reservation style is allowed. Furthermore, if a PLR node
receives a path message with fast-reroute requested with facility method and the FF
reservation style, it will reject the reservation. The one-to-one detour method supports
both FF and SE styles.
RSVP Message Pacing
When a flood of signaling messages arrive because of topology changes in the network, signaling
messages can be dropped which results in longer set up times for LSPs. RSVP message pacing
controls the transmission rate for RSVP messages, allowing the messages to be sent in timed
intervals. Pacing reduces the number of dropped messages that can occur from bursts of signaling
messages in large networks.
To Next Page
Loading...
