On Jan 30, 2008 9:41 PM, Todd Underwood <todd-nanog@renesys.com> wrote:
On Thu, Jan 31, 2008 at 01:56:42AM +0000, Paul Ferguson wrote:
For what its worth, Todd Underwood has a very good overview of the countries affected by this outage over on the Renesys Blog here:
http://www.renesys.com/blog/2008/01/mediterranean_cable_break.shtml
while i very much appreciate the compliment, this work was all done by my colleagues at renesys earl zmijewski and alin popescu. i've been following the routing events around this cable break, though.
there are some interesting findings here about who (what carriers, what countries) were critically dependant on these cable systems.
In the Med/IO cable case, a ship dropped an anchor on the cable, something that is 1:1,000,000 shot, but happens. At least they know where it is. The failure to contract the maintenance ship tighter on a route that turns out to be "that vulnerable" is probably of concer for users of that cable now as well. A lot of the impact is likely also due to people not buying protect circuits or bothering to understand the IP architecture. That is something that is becoming common globally, IMHO. Folks assume that IP will route around the damage. Sure it will, if all the physical layer paths aren't busted. Layer 1 really does "rock". Watching BGP announcements seems "less important" in these erious performance impacting cases, to me, than understanding the underlying architecture and what the root cause a half step above the anchor and a half a step below the advertisement was. Looking forward to Rod Beck's response. :-) Best, Marty