Feature #3143
openDiscover RIB management prefix of connected gateway router
0%
Description
Design a procedure to discover the RibMgmt prefix of a connected gateway router.
Scenario:
- An end host is already connected to zero or more gateway routers. For each router, the end host knows a prefix to access its NFD-RIB.
- The end host connects to one or more additional gateway routers.
- For each new router, the end host should discover a prefix to access its NFD-RIB.
Requirements and assumptions:
- The end host must be able to verify responses from gateway routers against a trust schema.
- The end host should support multiple simultaneous discoveries, because it's possible that an additional connection is established before the previous discovery is complete.
- The end host should tolerate multiple connected gateway routers having the same NFD-RIB prefix.
- It's assumed that the NFD-RIB prefix of a connected router does not change without breaking the connection.
Updated by Junxiao Shi about 9 years ago
- Related to Feature #3142: Configurable top-level prefixes for NFD-RIB added
Updated by Junxiao Shi about 9 years ago
- Blocks Feature #3145: Collect and publish routable-prefixes added
Updated by Muhammad Hosain Abdullahi Sabet almost 8 years ago
3.For each new router, the end host should discover a prefix to access its NFD-RIB.
AND
- The end host should tolerate multiple connected gateway routers having the same NFD-RIB prefix.
So you are certain that "a prefix" in "3." of scenario section will not be a unique one, otherwise req. no 3 will not exist, right? Thus, maybe -imho- we could have a mapping-like phase in end-host's nfd side which maps prefixes to faces and vice versa?
Updated by Junxiao Shi almost 8 years ago
The end host should tolerate multiple connected gateway routers having the same NFD-RIB prefix.
The discovery protocol only needs to find out what prefix each gateway router is using. In case of multiple routers having the same NFD-RIB prefix, further operations may not run smoothly, but it's not discovery protocol's concern.
a mapping-like phase in end-host's nfd side which maps prefixes to faces and vice versa
Yes, the outcome of this protocol is to have a mapping between face and NFD-RIB prefix, as indicated in scenario step1.