North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: [[email protected]: Re: NapsRus]

  • From: Randy Bush
  • Date: Fri Oct 18 22:36:25 1996

We also have NAPs with no traffic, unless you count the Email about them,
of course.

But, given what I have seen of driving in your area, this is not the case
for the VA ZPE-7392 NAP, even with no customers.

randy
- - - - - - - - - - - - - - - - -

o renumber.
>> I would expect the world to see the following announcements.
>> A - I announcing a shorter prefix, A
>> B - P announcing a more specific for the /24 allocated to C
>> Is this not what we all expect?
>> Why I ask is we have provider P telling I that they should stop announcing A
>> and instead announce a bunch of prefixes around the /24.
> Let's get the whole picture:
>   before the move:
> 	P announces   A'/L1 to <the world>
> 	B announces   A/L2  to P  (where L2 > L1)
> 	C announces   B/24  to B

You have changed notation and possibly misunderstood what is happening.  To
retain the original notation (Hints: Provider, ISP, Customer; P and I are
multihomed, C never is),

  before the move:
     P announces A'/L1 some block containing I's block, A
     I announces A/L2 (where A is within A' & L2>L1) to P and other(s)
     P and others announce A/L2 which they hear from I
     B/24 is contained in A and is a static route allocated to C and only
       known internally to I

> There is no reason for P to announce anything more specific than A'/L1
> to the rest of the world unless B is multihomed.

I (which I think you renamed B) is multihomed.

>   after the move [(if B is multihomed)]:
> 	P announces   A'/L1 to <the world>
> 	B announces   A/L2  to P  (where L2 > L1, L2 < 24)
> 	B announces   A/L2  to <other provider O>
> 	P announces   A/L2  to <the world>
> 	O announces   A/L2  to <the world>
> 	C announces   B/24  to B
> 	P announces   B/24  to <the world>
> 	P announces   A'/L1 to <the world>
> 	B announces   A/L2  to P  (where L2 > L1, L2 < 24)
> 	C announces   B/24  to B

This assumes C is also multi-homed, which is not the case.  They merely
changed provider from I to P.  The C also stands for Churn :-).

  after the move (as it should be):
     P announces A'/L1 some block containing I's block, A
     I announces A/L2  (where A is within A' & L2>L1) to P and other(s)
     P and others announce A/L2 which they hear from I
     B/24 is contained in A and is a static route to C and known internally
       to P (note change from I)
     P should announce B/24

  after then P claims that the following must occur:
     P will not do the last above, announce B/24
     I is being told to announce a *mess* of *pieces* of A (to 'get around'
       B/24) to P and their other upstream(s) because P can not seem to
       properly announce all of A', A, and B
     P and others should announce the *many* *pieces* of A/L2 they hear
       from I
     P still announces A', which is now the only covering prefix for B/24,
       thereby turning a /24 into many smallish announcements.  

And, given prefix length filters around the net, guess who eats it, I and
I's customers who now have many pieces of A as opposed to A.  And this
gives one a suspicion why P and C don't want B/24 to be announced.  But why
should I, I's customers, and the rest of the net pay for this?

> The last part is unfortunate.

I would call it seemingly inept (I am still trying to understand why P
can't just do the right thing), clearly asocial (though not holding a
candle to Telia, Electric Lightwave, and the other shining stars of Tony's
list, but I am not bing asked to be an accessory to those <bleep>s), and
sufficiently embarrassing that I have not named P (and no, it is not SL).
I am just trying to be sure this is indeed as stoopid as it appears to me
before making a bit (more) of a fuss.

Oh clueful ones at P, please explain where I am misunderstanding things as
usual.

randy
- - - - - - - - - - - - - - - - -