As to the earlier discussion of 'sophisticated threats' and traffic
injections.
An old acquaintace with good amount of scar tissue recently pointed out
to me
a most instructive paper in another areathat interestingly enough
documents what I wrote half in jest a while earlier.
The read is 'B4: Experience with a Globally-Deployed ...' by U. Hoelzle
& rest of Google gang in Aug. SigComm (for the faint of heart, it's an
easy read, light on math, rich on practical experience).
Section 7 documents a ('classical' in my experience) routing control
melt-down as to the ones I experienced over years.
Starting with second most common problem (human configuration error)
leading right into most common problem (i.e. naive implementation not
priotizing lsp/hellos) leading to the (in this case not catastrophical
since when you own the infrastructure of all routers you can reboot them
all) total meltdown (kind of, TE stayed up in frozen state).
The 'conclusions' in the paper are correctwhen also some a tad belated
['We need to test things under load' ;-) ]
Again, entertaining read for all practictioners of the routing game and
in itself quite impressive workin terms of size/volume & novelty of
approach (for the specific use case);-)
--- tony
_______________________________________________
OSPF mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ospf
