Re: How is a "route" defined?

[Reinaldo Penno Filho <[e-mail]>]

On Thu, 11 Dec 1997, Ken Rosenfeld wrote:

vc>In the context of route dampening, how is a "route" defined?
vc>
vc><draft-ietf-idr-route-damp-00> states (p. 13):
vc>  "A route here is considered to be a tuple containing at least
vc>   NLRI prefix, next hop, and AS path."
vc>


Hi, 

Sorry if this is a newbie question but I´m having a hard time
understanding the difference between a Route, NLRI and their relation
to UPDATE messages.

In the <draft-ietf-idr-bgp4-06.txt> is said:

-->   For purposes of this protocol a route is defined as a unit of
   information that pairs **a** destination with the attributes of a path
   that destination:   <--

-->An UPDATE message is used to advertise a *single* feasible route<--


Does "a destination" means one AS or NLRI<->path attribute information?

Cause if a destination is one NLRI seems to me that a UPDATE message
can advertise several routes..

Or..

Does this mean that a collection of NLRIs and their path attributes 
is a *single* route, even tough some NLRI<-> path attributes tuples
advertise different destinations (networks)? 

Something like this...

NLRI 1 --\
          \
NLRI 2 ---- Path Attributes
          /
NLRI 3 --/


(NLRI 1, Path At.)
(NLRI 2, Path At.)
(NLRI 3, Path At.)

A many to one relationship. 


So another update message to the same AS with the same NLRI but other path
attribute is another route? 

Sorry for the long and possible intricate message...

Best Regards,

Reinaldo
      
----------------------------------------------------------------------------
Brazilian Reserch Network
http://www.rnp.br
[e-mail]
----------------------------------------------------------------------------