At 05:30 PM 9/2/2008, Yakov Rekhter wrote:
Here is the text I'll add to the spec to discuss this issue:
Use of Traffic Engineering Attribute does not increase the number
of routes.
When the routes differ in other than the Traffic Engineering
Attribute (e.g., differ in the set of Route Targets, and/or
NEXT_HOP), use of Traffic Engineering Attribute has no impact
on the number of BGP Update messages required to carry the routes.
Only when the routes have all, but the Traffic Engineering
Attribute the same, use of the attribute may increase the number
of BGP Update messages required to carry the routes.
Yakov.
Is this from babel fish? How about: Use of Traffic Engineering Attribute does not increase the number of routes, but may increase the number of BGP Update messages required to distribute the routes. When the routes differ in other than the Traffic Engineering Attribute (e.g., differ in the set of Route Targets, and/or NEXT_HOP), use of Traffic Engineering Attribute has no impact on the number of BGP Update messages required to carry the routes. There is also no impact when routes share all other attribute information and have an aggregated or identical Traffic Engineering Attribute. When routes share all other attribute information and have different Traffic Engineering Attributes, routes must be distributed in per-route BGP Update messages rather than a single message. Lou