File tree Expand file tree Collapse file tree
examples/bgpv4/BgpIpv6Basic Expand file tree Collapse file tree Original file line number Diff line number Diff line change 1+ # BgpIpv6Basic — BGP over IPv6 (MP-BGP, RFC 4760)
2+
3+ A minimal IPv6 EBGP example: two single-router autonomous systems peer over IPv6 and each
4+ advertises the IPv6 network behind it.
5+
6+ ```
7+ hostA --- rA (AS 65001) === rB (AS 65002) --- hostB
8+ 2001:db8:0:10::/64 2001:db8:0:1::/64 2001:db8:0:20::/64
9+ ```
10+
11+ What it demonstrates / verifies:
12+
13+ - A BGP session is established over ** IPv6 TCP** (the ` Bgp ` module runs with
14+ ` addressFamily = "ipv6" ` and drives the IPv6 routing table).
15+ - IPv6 reachability is exchanged using the ** ` MP_REACH_NLRI ` ** path attribute
16+ (AFI=2 / SAFI=1), per RFC 4760 — rA advertises ` 2001:db8:0:10::/64 ` , rB advertises
17+ ` 2001:db8:0:20::/64 ` , each via a ` <Network> ` element in ` BgpConfig6.xml ` .
18+ - The learned routes are actually ** installed and used** : the configurator adds only
19+ on-link routes (` addStaticRoutes = false ` + manual ` <route> ` entries), so the inter-AS
20+ routes can come * only* from BGP. ` hostA ` pinging ` hostB ` therefore succeeds only once
21+ BGP has converged.
22+
23+ Notes:
24+
25+ - The BGP Identifier is a 4-octet router id even for IPv6 (RFC 4271/6286); it is set
26+ explicitly via ` bgp.routerId ` since an IPv6 router has no IPv4 address to derive it from.
27+ - The manual on-link ` <route> ` entries are a workaround for a configurator limitation
28+ (` addDirectRoutes ` is only applied when ` addStaticRoutes ` is true); see
29+ ` plan/pending/configurator-decouple-direct-routes.md ` .
You can’t perform that action at this time.
0 commit comments