Ouroboros multicast conjecture

From Ouroboros
Jump to navigation Jump to search

The Ouroboros multicast conjecture (apologies, but we had to give it some name) states that unicast is not equal to 1:1 multicast. Similarly, it is not possible to make unicast or multicast/broadcast transparent to the entity that directly engages in the unicast or multicast/broadcast communication. Or, in yet another formulation, it is not possible to have a single transparent API for multicast/broadcast and unicast.

This seems like a rather trivial fact, but we didn't find this stated explicitly in literature.

The current formulation is rather intuitive, it would be interesting to have an adequate precise description of this conjecture, preferably with a formal proof.

Origin

We came to this conjecture when starting to implement multicast in the prototype, starting from the whatevercast concept proposed in RINA. During the implementation, it became clear that to implement it, we were re-implementing a lot of concepts that were already present for unicast layers, most particularly enrollment and the code for disseminating link-state routing packets. It also became clear that we could not devise an API that somehow was oblivious to the destination being a single unicast node or a (dynamic) group of nodes, in other words: there is a distinction between 1:1 multicast and unicast. Intuitively, when starting a unicast flow, there is no way to add new nodes to that conversation; it has to start as a different kind of flow with certain limitations on authentication, encryption and QoS characteristics.

This led us to conclude that multicast is actually a process that consists of two phases:

  1. The creation of a network (and nodes can join and leave at will)
  2. Broadcasting the packets on that network

We call this network a Broadcast Layer.

So, given that insight, we had the option to either

  1. use different function calls towards unicast and multicast, or
  2. use the same function call, but make a distinction between unicast and multicast names in some namespace.

We opted for the former (the Unicast API and Broadcast API), but in any case, the application "has to know" whether it is doing multicast or unicast when it uses the API.

Examples

IP networks specify multicast/broadcast using a specific reserved range in their address spaces. There exist protocols to create a tree between members of an IP multicast group. According to the Ouroboros model, this multicast tree corresponds with a Broadcast Layer, and IGMP is its enrollment protocol. An IP multicast address corresponds with the Broadcast Layer name.

Notes

This conjecture does not imply that an application has to be aware of the use of multicast/broadcast communication in further down the network/stack, such as IP unicast implemented over Ethernet may leverage broadcast as used by the Address Resolution Protocol (ARP).