docs: add doc for unnumberd BGP

Signed-off-by: ISHIDA Wataru <ishida.wataru@lab.ntt.co.jp>

1 files changed, 94 insertions, 0 deletions

diff --git a/docs/sources/unnumbered-bgp.md b/docs/sources/unnumbered-bgp.md
new file mode 100644
index 00000000..0fa6b9fa
--- /dev/null
+++ b/docs/sources/unnumbered-bgp.md
@@ -0,0 +1,94 @@
+# Unnumbered BGP
+
+BGP is not only for the Internet. Due to proven scalability and configuration
+flexibility, large data center operators are using BGP for thier data center
+networking [[draft-ietf-rtgwg-bgp-routing-large-dc](https://www.ietf.org/id/draft-ietf-rtgwg-bgp-routing-large-dc-11.txt)].
+
+In typical case, the topology of the network is CLOS network which can offer
+multiple ECMP for ToR switches.
+Each ToR switches run BGP daemon and peer to uplink switches connected with
+P2P link.
+
+In this case, since all switches are operated by single administrator and trusted,
+we can skip tedius neighbor configurations like specifing neighbor address or
+neighbor AS number by using unnumberd BGP feature.
+
+Unnumbered BGP utilizes IPv6 link local address to automatically decide who
+to connect. Also, when using unnumberd BGP, you don't need to specify neighbor AS number.
+GoBGP will accept any AS number in the neighbor's open message.
+
+## Prerequisites
+
+To use unnumbered BGP feature, be sure the link between two BGP daemons is P2P
+and IPv6 is enabled on interfaces connected to the link.
+
+Also, check neighbor's IPv6 link local address is on the linux's neighbor table.
+
+```bash
+$ ip -6 neigh show
+fe80::42:acff:fe11:5 dev eth0 lladdr 02:42:ac:11:00:05 REACHABLE
+```
+
+If neighbor's address doesn't exist, easiest way to fill the table is `ping6`.
+Try the command below
+
+```bash
+$ ping6 -c 1 ff02::1%eth0
+PING ff02::1%eth0 (ff02::1%eth0): 56 data bytes
+64 bytes from fe80::42:acff:fe11:5%eth0: icmp_seq=0 ttl=64 time=0.312 ms
+--- ff02::1%eth0 ping statistics ---
+1 packets transmitted, 1 packets received, 0% packet loss
+round-trip min/avg/max/stddev = 0.312/0.312/0.312/0.000 ms
+```
+
+More reliable method is to run [radvd](http://www.litech.org/radvd/) or 
+[zebra](http://www.nongnu.org/quagga/) to periodically send router
+advertisement.
+
+## Configuration via configuration file
+
+```toml
+[global.config]
+  as = 64512
+  router-id = "192.168.255.1"
+
+[[neighbors]]
+  [neighbors.config]
+    neighbor-interface = "eth0"
+```
+
+## Configuration via CLI
+
+```bash
+$ gobgp global as 64512 router-id 192.168.255.1
+$ gobgp neighbor add interface eth0
+$ gobgp neighbor eth0
+BGP neighbor is fe80::42:acff:fe11:3%eth0, remote AS 65001
+  BGP version 4, remote router ID 192.168.0.2
+  BGP state = BGP_FSM_ESTABLISHED, up for 00:00:07
+  BGP OutQ = 0, Flops = 0
+  Hold time is 90, keepalive interval is 30 seconds
+  Configured hold time is 90, keepalive interval is 30 seconds
+  Neighbor capabilities:
+    multi-protocol:
+        ipv4-unicast:   advertised and received
+        ipv6-unicast:   advertised and received
+    route-refresh:      advertised and received
+    extended-nexthop:   advertised and received
+        Local:  nlri: ipv4-unicast, nexthop: ipv6
+        Remote: nlri: ipv4-unicast, nexthop: ipv6
+    four-octet-as:      advertised and received
+  Message statistics:
+                         Sent       Rcvd
+    Opens:                  1          1
+    Notifications:          0          0
+    Updates:                1          0
+    Keepalives:             1          1
+    Route Refesh:           0          0
+    Discarded:              0          0
+    Total:                  3          2
+  Route statistics:
+    Advertised:             1
+    Received:               0
+    Accepted:               0
+```