> much talk and fud about this. in practice, this does not seem to be
> a problem.
It is clear that there is a *potential* for trouble. If one controls
the environment sufficiently, uses protocols that are relatively
robust to network failures, one can make it work. Even work quite well
it would seem. That is fine and we should acknowledge this.
But it is also not helpful to just assert "it works in practice"
without acknowledging that there are real dangers when used
inappropriately. i.e., the "it" needs to be qualified, as not all
"its" will work equally well.
> e.g. some years back, we did a two hour streaming videocast to over
> 100k users, 200mb of traffic, from a dozen anycast servers. not one
> noc call.
If the metric of "it works" is defined by "how many calls to the NOC",
we're in pretty bad shape. My experience in calling various NOCs is
that it is a waste of time. Only bother when things are totally broken
and they haven't fixed themselves after a prolonged period of
time. I.e., call the NOC only after exhausting all other
alternatives. YMMV.
Thomas
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
