|Still from Renesys animation|
The Internet senses its own failures in two ways, automatically and in cooperation with humans.
A network is removed from the global routing table a few seconds after it goes down. Renesys tracks the dynamic state of the Internet by monitoring that table.
This animation shows the percent of networks that are down in small geographic areas hit by the storm. Dark green indicates that at least 99.95% of the networks are up and the dark red indicates that more than 5% are unreachable.
They report that in Manhattan the typical outage rate is around 10% and point out that "silencing ten percent of the networks in the New York area is like taking out an entire country the size of Austria, in terms of impact on the global routing table." That is the bad news. The surprisingly good news is that 90% of the data centers are still up -- running on backup diesel power and caffeine.
On the right, we see another Renesys view showing network outage by state over time. As we see, New York was the hardest hit with around 1,200 networks off line at the peak, but some have come back on-line.
Renesys monitoring is automatic, but people are also monitoring the network with the aid of tools like Twitter. Andy Carvin is known for his use of Twitter and other Internet tools in producing real time news reporting (see the presentation at this location) and the same approach has been used in disaster reporting.
Using Twitter and other sources, Rich Miller of the Data Center Knowledge blog has been reporting on data center outages
Finally, let's note that the human-Internet collaboration on disaster reporting or any other task was anticipated long ago by J. C. R. Licklider who, in the 1960s, wrote of man-computer symbiosis, envisioned the Internet and was instrumental in funding much of the research that led to the Internet and modern personal computers. (Read two of his highly influential papers here).