question about reservation IDs

as i mentioned in an earlier message to the list, i am creating a thing 
called a reservation set that is used for 1:n and m:n recovery purposes.

- i think this is only useful for recovery purposes, does anyone else 
think it might be useful for anything else and therefore should be 
generalized?

- (more important question)  i have been going back and forth between 
thinking these sets should occupy their own reservation number space.  
they are different in that they really don't tie up resources, but 
rather just aggregate previously reserved resources and perhaps should 
not count against the max reservations limit.

first thoughts were to take a bit away from the reservation id and 
define an aggregation (A) flag
0- individual reservation (backward compatible)
1- set reservation

then i decided it is silly, there is no driving reason to make them a 
separate space - there won't be so many compared with actual 
reservations that it woul make much of a dent in max reservation count.
but it keeps nagging at me that i should.  so i go back and forth.

also i don't currently plan to create a new delete reservation message. 
  rather i think it just has different behavior, if it is an aggregate 
it deletes the set, but not the reservations themselves.  i am keying 
this off of the A flag but it really doesn't matter since if they 
occupy the same reservation space i think we can expect the switch to 
know which are which.

any opinions?

a.