RE: [RRG] Which Side to Control Ingress Link Selection?

[<hannu.flinck@nsn.com>]

Christian

Incentive wise the one who carries the cost of the use of a resource
should have the choice of setting the policies and control the usage.
Receiving side knows better the local conditions of the CPE network
behind it than the remote end. The remote, or sending side, doesn't have
much visibility to the local link states and the capabilities or load
situation of the CE equipment. And I think that such details should be
abstracted away from the sending side. It should suffice, that the
receiving side announces its reachability and preferences between its
multiple egress points if such existed. 


Regards Hannu     

>-----Original Message-----
>From: ext Christian Vogt [mailto:christian.vogt@nomadiclab.com] 
>Sent: Friday, April 18, 2008 12:05
>To: Flinck Hannu (NSN - FI/Espoo)
>Cc: rrg@psg.com
>Subject: Re: [RRG] Which Side to Control Ingress Link Selection?
>
>Hannu,
>
>you are too far ahead. :-)  I was still thinking on the 
>incentives/ economic level.  Whether we decide that an ingress 
>link should be selected by the receiving side or by the 
>sending side, I am sure we will eventually find a good 
>technique for it.  But let's first decide which approach would 
>be preferable.  (Feedback from the receiving side, as you 
>mention, will certainly be necessary if we decide that the 
>receiving side should select its ingress links.)
>
>But you seem to be in favor of letting the receiving side 
>select.  Is this right?  And if yes, please elaborate what 
>makes you think so.
>
>- Christian
>
>
>
>hannu.flinck@nsn.com wrote:
>>  Hello
>> 
>> I am wondering what triggers the egress indirection router 
>to require 
>> fresh reachability information for a receiving edge? Doesn't this 
>> imply that the receiving edge indicates the need somehow?
>> 
>> All this sounds quite complicated and fragile if 
>communication between 
>> ingress and egress and mapping system is needed to accomplish 
>> something that is already working through BGP.
>> 
>> Why not just manage the tunnel end points between egress and ingress 
>> routers? If either of the tunnel end points want to change 
>local their 
>> end point (in an managed way, not due to a failure), they 
>will signal 
>> to their intention to the other end and let the remote side 
>know what 
>> would be the "next" tunnel end point.
>> 
>> Maybe I am missing something here but how could the sending 
>side know 
>> better what the receiving side wants?
>> 
>> 
>> Regards Hannu   
>
>
>
>

--
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