[Sis-ams] comment on consistency

[Scott Burleigh]

Ray, Timothy J. (GSFC-583.0) wrote:
> Dear WG members,
> 
> This email concerns an inconsistency in the Meta-AMS protocol.  The 
> inconsistency does not cause any technical problem with the protocol, 
> and I’m not requesting that we change anything.  However, it seems 
> worthwhile to bring up the inconsistency and see if anyone else cares to 
> comment.
> 
> The way a registrar registers is somewhat different than the way that a 
> node registers.  If the node’s Node_registration message is accepted, 
> then it automatically receives info about its peers (it receives an 
> I_am_here message for each peer node).  If the registrar’s 
> Announce_registrar message is accepted, it does not automatically 
> receive info about its peers (it has to send a Msg_space_query message 
> to get the information).

I think it's important not to get carried away and start over-polishing the specs, but this is a small simplification that I think makes sense.  All that would be necessary, I think, is for the configuration server to respond to announce_registrar with both registrar_noted (to terminate the new registrar's configuration server interrogation cycle) and also -- without further prompting -- the cell_spec MPDUs for all cells of the message space.  [In which case I believe there's no further need for msg_space_query.]  It should make implementations and the spec marginally simpler.  Does anyone see a downside to adopting this change?

Scott