FRRouting Release 10.7.0 [Download]

FRRouting Release 10.7.0 Available for Download

We are pleased to announce FRR release 10.7.0.

FRR 10.7.0 brings a long list of enhancements and fixes with 1433 commits from 105 developers.

Thanks to all contributors!

Debian Packages - https://deb.frrouting.org

RPM Packages - https://rpm.frrouting.org

Docker - quay.io/frrouting/frr

Release Overview

New Features Highlight

• BFD authentication with keychain support
  • BFD sessions can now use the FRR keychain framework for authentication.
  • Supported algorithms include cleartext, SHA-1, and the meticulous variant required by operators running strict authentication policies.
  • Keychains integrate with BFD profiles and northbound configuration for single-hop, multi-hop, and SBFD sessions.
• BGP-LS SRv6 extensions
  • BGP Link-State now advertises and processes SRv6 topology information, including:
    • SRv6 Capabilities TLV
    • SRv6 Locator TLV
    • SRv6 SID NLRI (including End, End.X, and LAN End.X forms)
    • SRv6 Endpoint Behavior TLV
    • SRv6 SID Structure TLV
  • Static SRv6 SIDs configured in zebra can be exported into BGP-LS, enabling SRv6-aware TE controllers to consume locator and SID information from FRR.
• BGP route-map based allowas-in
  • allowas-in can be applied selectively through route-maps instead of only as a blanket peer or address-family knob.
  • This allows finer control over which received routes may contain the local AS in the AS_PATH, improving policy flexibility in complex multihomed and hub-and-spoke designs.
• OSPFv2 RFC 4222 control-plane DSCP marking
  • OSPFv2 can classify control packets into high- and low-priority queues per RFC 4222 recommendation 1.
  • Recommendation 2 extends the neighbor inactivity timer behavior for low-priority control traffic.
  • Operators can mark OSPF control traffic with DSCP values to protect routing protocol stability under congestion.
• OSPF BFD quick neighbor
  • When BFD is enabled on OSPF interfaces, a quick-neighbor mode reduces the time needed to bring up OSPF adjacencies over BFD-monitored links.
  • This is useful on fabrics where fast reconvergence after link-up is required.
• PIM IGMP proxy route-map filtering
  • IGMP proxy joins and prunes can be filtered with ip igmp proxy route-map.
  • A new route-map match condition, match multicast-source-interface, allows policy based on the interface where a join was learned.
  • This helps deployments that proxy IGMP between access and core segments while restricting which groups or sources are forwarded.
• PIM dense-mode Assert on multi-access LAN
  • Dense-mode wrong-interface traffic on shared LANs now runs the RFC 3973 Assert procedure instead of immediately pruning all neighbors.
  • On point-to-point links the previous immediate-prune behavior is preserved.
  • This elects a single forwarder on multi-access segments and avoids pruning active branches when duplicate (S,G) packets arrive on a non-RPF interface.
• PIM Auto-RP allow-rp support
  • A new allow-rp configuration accepts Auto-RP groups even when the embedded RP address does not match the local RP configuration.
  • IPv4 and IPv6 Auto-RP are supported through northbound/YANG modeling.
  • This helps mixed-vendor or transitional Auto-RP deployments where strict RP identity checking would otherwise drop valid control traffic.
• Static route per-route metric and nexthop ECMP weight
  • Static routes now support a per-route metric as a non-key attribute, independent of the existing administrative-distance key.
  • Static ECMP nexthops can carry weights for unequal-cost multi-path forwarding.
  • Together these align static routing policy more closely with zebra’s nexthop-group handling used by other protocols.

Full Changelog: https://github.com/FRRouting/frr/compare/frr-10.7.0-dev...frr-10.7.0