> From: "Tony Hain" <alh-ietf(_at_)tndh(_dot_)net>
>> Since there's no technical difference between PI and number
>> portability, I expect approval of PI-space will lead to portability
>> anyway.
>>
>> Yes, the current criteria for PI-space are rather limited, but since
>> there's no particular technical rationale for picking /N versus /M, I
>> expect to see a salami-slicing political debate in which people will
>> demand smaller and > smaller blocks be supported, because to do
>> anything else is dumping on the "little guy", while letting the "big
>> players" have acess to something the small players don't.
> I happen to agree with you here, which is why I have been advocating a
> particular geo approach ... be scaled up and down as far as necessary
> to contain the routes.
> ... the IESG has not understood the necessity to have a working group
> to refine a globally acceptable approach.
Perhaps I'm not understanding your point, but the whole point of my comment
is that there's no "bright technical line" differentiation between one size
and another, and we'll inevitably wind up being forced to support the
smallest possible one.
Perhaps the IESG's reluctance is because this now turns into something like
the old (PC-incorrect) joke about the gentleman and the debutante in the
railway - we've already decided what we are, now we're only haggling about
the price (or block-size, as the case may be).
>> Sigh, we're going to be paying the price for not (long ago) setting
>> up a charging system for having a route be visible in the
>> "default-free zone".
> The existence of a "default-free-zone" was long ago relegated to myth
> status. There are different views of what this mythical entity
> contains, so by definition there is no single zone.
I am fully aware that for a variety of reasons, including policy
limitations, as well as different aggregation action boundaries (i.e. places
where routes to X.1 ... X.n get discarded in favour of a single route to X),
different in the 'core' (to use another somewhat nebulous term) routers will
have somewhat different routing tables.
My reference was more in the sense of 'routers which do not have a routing
table which consists of a small number of destinations, and use a default
for the rest' - i.e. routers which *do* have to have routing information for
the vast majority of the destinations in the entire network.
If we had a cost-structure which actually reflected reality (so that
PI-address X, which has to have a routing table entry in every router which
wants to be able to send traffic to it, had to pay for those routing table
entries), we might have a little less enthusiasm for the PI-approach.
PI is like spam - it looks attractive to the people using it, because it's
free to them. The fact that it costs *other* people money is something
they don't care about - it's not coming out of their pocket.
Noel
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf