John,

To reply to my own message, after reading through the rest of this chain. 

Is all we're trying to do here is to establish a "custodial chain" of a route 
to prevent some ill-behaving AS in the middle attempting to hijack a route, 
effectively by pretending that the source AS is behind it, since we're 
origin-validating, and hoping that its neighbors end up preferring it's 
longer-AS announcement over other, likely preferred routes, from legitimate 
sources?

If that is the case, having a set of policy objects expressing AS relationship 
should do the same thing  and more with less overhead? (yes, I know that data 
integrity becomes an issue, but data integrity is always an issue.)

Andrew

On Feb 28, 2011, at 8:16 PM, Andrew Lange wrote:

> John,
> 
> But wouldn't a record of an existing announcement also show that AS_B did in 
> fact announce which of AS_A's routes and in what form?  Why does it need to 
> be signed if all we want to do is record what happened?  Perhaps I'm missing 
> something....
> 
> Andrew
> 
> On Feb 24, 2011, at 5:27 AM, John G. Scudder wrote:
> 
>> Andrew,
>> 
>> On Feb 24, 2011, at 1:17 AM, Andrew Lange wrote:
>>> Given the thread, I can understand your frustration.  And, I could have 
>>> made myself more clear.  I'll try:  given the policies that AS_B might be 
>>> implementing, we cannot know for certain, without AS_B publishing their 
>>> policies, if AS_B should in fact be announcing which of AS_A's routes or in 
>>> what form, hence the use of the term "Plausible." 
>> 
>> Assume a model where the AS path is signed such that the recipient is able 
>> to verify that for each arc in the path, the further AS did sign it to the 
>> closer AS.  To take your example and extend it, suppose we have
>> 
>>      AS_A --> AS_B --> AS_C --> AS_D
>> 
>> (A announces a route to B announces it to C announces it to D.)  If D is 
>> able to verify the signature on the (B, C) arc, then it knows (for some 
>> value of "knows" of course :-) that B did actually send the route to C.  
>> Whether that fulfills "AS_B should in fact be announcing which of AS_A's 
>> routes or in what form" I leave it to you to interpret -- it doesn't protect 
>> against a policy misconfiguration on the announcing router of AS_B.  An 
>> abstract representation of AS_B's policy could do that, but functionality 
>> like that is beyond the scope of the proposed charter item.  The fact that 
>> systems to provide exactly this kind of abstract policy publication have 
>> fallen into disrepute doesn't fill me with optimism about this approach, 
>> despite my own personal history with it.
>> 
>> I interpret the proposed charter item to be asking not "if AS_B *should* in 
>> fact be announcing which of AS_A's routes or in what form" but rather 
>> roughly "if AS_B *did* in fact announce which of AS_A's routes and in what 
>> form".
>> 
>> --John
> 
> <smime.p7s>_______________________________________________
> sidr mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/sidr

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to