|
North American Network Operators Group Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical RE: Attacks Expose Telephone's Soft Underbelly
On Mon, 15 Oct 2001, Kevin Gannon wrote: > I guess this is really a question for Sean given your background. > Over on this side of the pond the 999 (112) service needs to be > 100% reliable/redundant, having seen a major CO melt down in > the capital I can atest that it works. > > My question is how is this achieved for both 999 services and > critical government services ? Surely buisnesses can learn > something from it ? Never watch sausage being made. The basic principles (aka best practices) are well understood. However, money is the driving factor in all decisions whether those decisions are made by the public or private sector. 9-1-1 (the US version of emergency number) service is very reliable, but has been disrupted the same things which disrupts telephone service in general. Not all public service answering points have redundant circuits. Not all end-offices have diverse paths. Even when redundant circuits exist, they've been groomed on to a common physical facility. Operator and software errors corrupt translation tables in switches. Much of the reliability comes not from preventing things from breaking, but by priority repair service when it breaks. Because 9-1-1 is usually repaired before most other services, it has the best MTBF/MTTR even if it breaks due to the same cause as other services. Of course, there is always the definition of working. If your phone doesn't work, you can't call 9-1-1, even if the PSAP is "working." Another problem in New York City was NYC's emergency operation center "bunker" was destroyed in the collapse of the world trade center tower. In addition to all the other problems, Verizon needed to install/re-route emergency circuits for almost everything connected to the EOC.
|