The EID is, by definition, always the IP address assigned to the host (e.g., 192.168.2.102 in the lesson). The EID remains static even if the host moves locations, while the RLOC changes dynamically to reflect the host’s new topological location.
In a sense yes. But the RLOC does not represent a next hop as a part of a series of next hops. It represented the location of the intended destination host. There is only one RLOC, whereas there are a whole series of next hops to get to the destination in traditional routing.
The role of a router as an ITR or an ETR depends on the context. For example, R1 is an ITR for Site 1 for all outgoing traffic. It is however an ETR for all of the addresses within Site 1. So it plays the role of both. For traffic going in the direction of Site 1–> Site 2, the ETR is R2 at Site 2, because those are the destination IP addresses we want to use. You can choose to configure one router to play both roles, or you can separate the roles to different devices.
I think I answered this question above. There is an ETR for each LISP site which informs the mapping database system of the EIDs that it serves.
Regarding the data plane itself, the UDP destination port is always 4341 when tunneling data between ITR/ETR and 4342 when performing EID registration? I’ve seen that both are used.
I don’t really understand what role the Map Resolver (MR) performs in LISP.
If MS is the main one that handles registrations and maintains mappings, why did they invent MR in the first place? Apparently, if the ITR does not have a mapping in the local cache, it asks the MR for it first.
MR also checks its local database and either replies or asks the MS.
The use of MR seems a bit redundant to me at the moment, maybe a bit unnecessary? If the MS contains all the necessary information, why should the ITR routers bother asking the MR device? What function does such a router actually perform? Does it also have any actual mappings or what is there for?
Yes indeed both ports are used, but each one for specific purposes. A destination port of 4341 is used for encapsulated data traffic between ITR and ETR. This is part of the data plane. When used, the source port is often ephemeral (dynamically chosen by sender). A destination port of 4342 is used for control plane communication, specifically for:
EID-to-RLOC registrations from ETR to MS.
Map-Request/Map-Reply messages between ITR/ETR
Map Resolver/Map Server.
More info about these ports and how they are used can be found at the LISP RFC (6830).
I understand your confusion, it’s something that confused me too when I was learning LISP!
The MR’s primary purpose is to handle mapping queries from ITRs and delegate them appropriately. When an ITR needs to resolve an EID and lacks a cached EID-to-RLOC mapping, it does not send the Map-Request directly to the MS. Instead, the request is sent to the MR. Acting as a kind of intermediary, the MR authenticates the sender, validates the format of the Map-Request, and then forwards it to the appropriate MS, which holds the authoritative mapping database.
The separation of roles, such as the authentication and direction of requests by the MR, and the maintenance of the authoritative mapping database by the MS, helps to make LISP scalable and expandable. It is somewhat similar to the hierarchical nature of the DNS system that has a hierarchy of domains.
