NEXUS DC · MPLS TE

Nexus: RSVP-TE tunnel with a dynamic path

IS-IS sends PE1’s traffic for PE2 over the short top path through P1, but those links can reserve only 50 Mbps. Build an MPLS TE tunnel that asks for 100 Mbps: CSPF must find the bottom path through P2 and P3, RSVP signals it, and autoroute announce puts PE1’s traffic into it. P1, P2, P3 and PE2 already run MPLS TE.

Nexus DC track · Cisco Nexus (NX-OS-style) · 6 objectives

Open in simulator More Nexus DC labs

Objectives

The simulator ticks each one as soon as the network meets it.

On PE1 enable feature mpls traffic-engineering, and on both core interfaces mpls traffic-eng tunnels and ip rsvp bandwidth (50000 toward P1, 1000000 toward P2).
Flood TE information in IS-IS: under router isis CORE, mpls traffic-eng router-id Loopback0 and mpls traffic-eng level-2.
Create interface tunnel-te 1 to 5.5.5.5, unnumbered to Loopback0, with bandwidth 100000, path-option 1 dynamic and no shutdown.
tunnel-te1 comes up through P2 and P3, because the top path can reserve only 50000 kbps (show mpls traffic-eng tunnels).
Add autoroute announce: PE1’s route to 5.5.5.5 now points to tunnel-te1 (show ip route isis).
Ping 5.5.5.5 from PE1: the echo crosses the tunnel.

Start this lab

Before you start

CSPF runs on the head end over the TE database that IS-IS floods; a switch or link that does not flood TE is invisible to it. The bandwidth request is a constraint, so CSPF skips links that cannot reserve it, even when the IGP prefers them. On NX-OS a tunnel-te interface starts shut down, so it needs no shutdown. Without autoroute announce the tunnel is up but carries nothing.

Concept

MPLS on NX-OS: a feature set, then features

NX-OS keeps MPLS behind a feature set, one step above ordinary features. The set is installed once and then enabled, and only then do feature mpls ldp, feature mpls traffic-engineering and feature mpls segment-routing exist. Each one in turn makes its own commands appear. The protocols themselves work as on any Cisco router: IS-IS carries the routes, TE data and SIDs, and labels are swapped hop by hop.

Concept

CSPF: the shortest path that meets the constraints

An IGP always takes the lowest-cost path. MPLS traffic engineering lets the head end choose a different one: IS-IS floods every TE link with its TE metric and reservable bandwidth, and CSPF (constrained shortest path first) runs Dijkstra only over the links that meet the tunnel’s constraints.

Concept

RSVP-TE: signalling the tunnel hop by hop

Once CSPF has a path, RSVP-TE builds the label-switched path along it. The head end sends a Path message carrying the explicit route and the bandwidth; the tail end answers with a Resv message that reserves the bandwidth and hands out labels on the way back.