[RRG] Abstraction of problem space

[Brian Dickson <briand@ca.afilias.info>]

I've been following the discussions on ID, locater, id-loc split, etc., 
with a lot of experience in the ISP address space/routing/aggregation field.

Here's my observation:

There is a reason that IPv4 PI space *is needed* along with PA space.
And, there is a reason that there was backlash against the original GSE 
proposal.
And, there is a reason that those issues won't and can't go away. They 
are one and the same.

Simply put, there is *not* a simple (strict) hierarchy in the 
provider/customer and peering ecosystem.
And, the "locator" and "identifier" in this abstract environment needs 
to be able to accommodate change.
Someone can move up, or down, the food chain, as it were, and establish 
new (or eliminate old) relations in that space.

And, specifically, an entity's apparent role needs to be able to be 
ambiguous, simultaneously being at more than one level in the food 
chain, and simultaneously having more than one "profile".
Especially when considered over a non-trivial period of time, i.e. the 
lifetime of the entity.  (By "entity" we can mean anything with a 
routing policy, be it an ASN, ISP, site, host, or whatever.)

I fear that too many ideas being discussed, even in relatively abstract 
terms, don't take this into consideration, or are optimized for 
instantaneous state ("snapshot").

And it is this aspect of the Internet that is particularly difficult to 
accommodate, since it fundamentally makes aggregation (and thus scaling) 
very difficult.

Basically, scaling requires "layering", e.g. aggregation, or 
indirection, or other similar techniques.
But, a dynamic location in the food chain, automatically means that we 
need the *ability* to routinely handle "layer violations".

I think drilling down on this, providing means to incorporate such 
violations *into* the addressing/identifier/locator/mapping system, is 
the way forward.
Not including this in any solution, is for us to be chasing our tails, 
or producing interesting recommendations that will not yield a solution 
that can be deployed universally.
Not handling (intrinsically, elegantly, scalably) layer "violations" 
means not meeting the requirements given to the RRG.

BTW - the ability to take a piece of IPv4 PA space, and start announcing 
it as if it were PI space, is an example of the kind of thing that needs 
to be supported.
Changing the role of a "piece" of the Internet, without renumbering.
This needs to be possible without knowing ahead of time, *which* piece 
of the Internet it will be, that does this.
It should be universally possible, and at the same time scale well, and 
(ideally) work on (some/many/most) existing (classes of) hardware.

IMHO.

Brian Dickson

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg