Re: MALLOC MIB

> I have addressed all but one of the comments of the "MIB 
> doctor", and the last comment made me think a bit.
> We have an "exclusion" table in the MIB right now, that
> I believe originated from AAP.  I'd like to propose removing
> the table.
> 
> As far as I can tell, nothing in the MALLOC architecture
> (RFC 2908) or the MADCAP protocol (RFC 2730) requires
> implementing any such concept, and those are what this MIB
> instruments.
> 
> As a reminder, the MIB currently contains 3 tables that are
> relevant to this question:
> Scope Table : contains the scope ranges known (e.g. 
> 		  Global, LinkLocal, IPv4 Local Scope, etc.)
> Allocation Range Table : contains the subranges legal
> 		  to allocate from (e.g. the piece of LinkLocal
> 		  Erik Guttman asked IANA for, all of IPv4 Local
> 		  Scope except the scope-relative addresses,
> 	        your own GLOP range, etc).
> Exclusion Table : "contains sub-ranges which are excluded
> 		  from being allocated".  The wording implies
> 		  these are subranges of the scope, disjoint
> 	   	  from subranges in the allocation range table.
> 		  The description also used to say entries may
> 		  be dynamically discovered via some protocol
> 		  such as AAP (which we already removed from the 
> 		  MIB).
> 
> I believe that the Exclusion Table should have been removed
> when AAP was removed, and that for today's uses, exclusions
> are unnecessary (since the Allocation Ranges are manually
> configured without needing to have a concept of exclusions, 
> or are acquired as part of experimental protocols outside 
> the scope of this MIB).
> 
> Any objections?  (Positive acknowledgements from any of 
> Mark, Steve, and Pavlin would be greatly appreciated :)

It seems to me that the allocatable address ranges can be expressed
only by the information in the "Allocation Range Table" instead of
(AllocRangeTable - ExclTable).
If we have both AllocRangeTable and ExclTable, it may be easier
for practical purpose to express the allocatable range(s); it may
also make simpler the interfacing with protocols/mechanisms that
originate/create those ranges (I guess the simplicity of such
interface was probably the primary motivation to have the
ExclTable added to the MIB).
However, my guess is that the MIB doesn't have to deal with
such practical issues. Further, having ExclTable adds the complexity
that anyone who reads the MIB MUST take into account the ExclTable
as well to figure-out the address ranges that are valid for
allocation.

Now that the AAP has been removed, I don't see a reason to keep the
Exclusion Table, so removing it is fine with me.

Pavlin

P.S. FWIW, MASC doesn't have the equivalent of excluding
sub-ranges.