Points of Failure (was Re: National infrastructure asset)

Even if diversity exists, we need to consider (from an engineering
perspective) capacity requirements.

Analogy example: FedEx, UPS and others provide huge quantities of shipping,
and sustaining the infrastructure growth of the Internet to some substantial
degree relies on this capacity.

What would the direct and indirect impact of losing the airport at Memphis,
or the FedEx sorting facility? Pretty substantial, and long lasting, I'd

Similarly, while certain peering locations and relationships have
redundancies, the Venn diagrams for union and intersection sets of bilateral
peering are non-trivial. It would take a large number of incident locations
to affect global route distribution substantially, but only two or three
incidents before the impact on available peering capacity (private and
public) would be noticable to the general end-user.

I think to some degree this should be taken seriously - we should all ask
ourselves, does our peering diversity address *capacity* issues sufficiently
to handle even one major site outage (eg the power outages at 25 Broadway)?
And in considering this, look at each site itself and what the impact of it
going away, even temporarily, would be.

(Our plans involve going to a large number of neutral colos, for just this
reason, among others.)

Brian Dickson
Velocita Corp.

this is interesting, published late last year.

Error and attack tolerance of complex networks



"We find that such networks display an unexpected
degree of robustness, the ability of their nodes to
communicate being unaffected even by unrealistically
high failure rates. However, error tolerance comes at
a high price in that these networks are extremely
vulnerable to attacks (that is, to the selection and
removal of a few nodes that play a vital role in
maintaining the network's connectivity). Such error
tolerance and attack vulnerability are generic
properties of communication networks."

--- "Dickson, Brian" <brian.dickson@velocita.com>