> Quoted material from document.
> >2. System-wide
> >Defining a single target for maximum convergence time for the real
> >Internet is absurd. As we mentioned earlier, the Internet is large
> >enough and diverse enough so that it is quite likely that new
> >changes are introduced somewhere before the system fully digests old
> >ones.
>
> Do we have an agreement on what convergence time means? In the
> current BMWG work, we have tried to make that definition more
> rigorous, and also recognize there are several kinds of convergence
> time with different scopes. These different times need distinct
> names to avoid confusion. While draft-ietf-bmwg-conterm-01.txt is
> officially focused on BGP, it really does try to come up with some
> fairly general terms.
I really agree that we need to have an agreement on what convergence
means. I have found myself having hard time discussing convergence with
people. As I see in BGP, there are at least 2 dimensions. The first
dimension is in terms of the size of the network. Is it the convergence
within a single router, or within an AS, or within the whole Internet? The
second dimension is in terms of the information that BGP carries.
Currently there are at least 2 important attributes that BGP carries for
each prefix, the ASPATH and the NEXT HOP. So does convergence mean that
routers have correct decision about the NEXT HOP, or the ASPATH, or both?
As far as forwarding behavior is concerned, routers need to know the right
NEXT HOP in order to forward packets correctly. This implies for a
customer connected to the Internet through a provider, it is enough for
its provider to converge since the customer will forward its packets to
its provider anyway, even though it still has wrong knowledge about the
global AS path. For reaching convergence about the AS path, it's a much
slow process. The Internet can probably be in a state that the AS path
never converges, while all forwarding can be done correctly.
Before agreeing upon these terms, it's really hard to set requirements on
convergence time.
/Yong
This archive was generated by hypermail 2b29 : Mon Aug 04 2003 - 04:10:04 EDT