Home · Documentation · Technology · MPLS-TE

MPLS Traffic Engineering — RSVP-TE, FRR, Bandwidth Constraints

Segment Routing reduced the need for state in the dataplane, but RSVP-TE remains the gold-standard for bandwidth-reserved, fast-rerouting LSPs in service provider transport. OcNOS implements the full RSVP-TE stack — CSPF, FRR link/node protection, DS-TE class-types, and auto-bandwidth — alongside SR-MPLS, on the same chassis.

RSVP-TE LSP with FRR Backup

A four-router ring carrying a primary RSVP-TE LSP from PE-1 to PE-3. The primary path (solid) is the CSPF-computed shortest path that meets the 1G bandwidth constraint; the FRR backup (dashed) protects the primary against a P-A link or node failure.

RSVP-TE LSP with Fast Reroute backup in a four-router ring Four-router MPLS ring. The primary RSVP-TE LSP follows the upper path with reserved bandwidth. A Fast Reroute (FRR) backup LSP follows the lower path, providing sub-50 ms protection on link or node failure. BW=1G BW=1G BW=1G FRR backup (link protect) PE-1 · headend CSPF · auto-bw P-A PLR / merge PE-3 · tailend RRO / RSVP P-B merge point P-C FRR detour RSVP-TE · CSPF · FRR <50 ms · DS-TE · AUTO-BW

Where MPLS-TE still wins

Segment Routing handles most modern transport elegantly, but there are still two things RSVP-TE does better: strict bandwidth admission control across the network, and Fast Reroute backups that pre-program a detour LSP at every PLR. For carriers with strict SLA contracts, capacity-managed wholesale services, or per-class-type DiffServ guarantees, OcNOS delivers a complete RSVP-TE stack that runs alongside (or instead of) SR-MPLS on the same hardware.

The OcNOS MPLS-TE implementation

RSVP-TE

Path / Resv signalling

Full RSVP-TE signalling with explicit-route (ERO), record-route (RRO), graceful restart, hello-based detection, and refresh reduction.

CSPF

Constraint-based path

CSPF using IGP-TE extensions (OSPF-TE / IS-IS-TE) — bandwidth, affinity, hop-count, SRLG-disjoint, and admin-group constraints.

FRR

Link / node protection

Facility-mode FRR with bypass tunnels at every PLR. Sub-50 ms repair on link, node, or SRLG failure with bandwidth-aware backup selection.

DS-TE

Class-types & bandwidth pools

Russian-doll and maximum-allocation bandwidth constraint models; per-class-type (CT0-CT7) reservations with strict admission control.

Auto-Bandwidth

Adaptive reservation

Auto-bandwidth resizing based on observed utilization with configurable adjust-interval and overflow / underflow thresholds.

Inter-AS TE

Per-domain expansion

Inter-AS TE LSPs with per-domain ERO expansion and PCE-based stitching for multi-AS service provider deployments.

What you get with OcNOS MPLS-TE

  • Coexists with SR-MPLS. Run RSVP-TE for legacy services and SR-MPLS for new ones on the same router — same IGP, same forwarding plane.
  • Carrier-grade scale. Tested with thousands of LSPs per node, including auto-bandwidth resize churn and large FRR backup meshes.
  • Operational tooling. CLI and gNMI streaming for LSP state, RSVP neighbour status, and bandwidth-pool utilization per class-type.
  • PCE-ready. PCEP support for stateful PCE deployments, including PCInitiate for controller-driven LSP placement.
SR-MPLS BGP-LU OcNOS-SP

Designing a TE-managed transport core? Talk to a network architect.

Request a Technical Demo →