Slashdot Mirror
Preparing for the Worst in IT
mplex writes "How vulnerable is the internet to terrorist attack? Is it robust enough to handle an outage on a massive scale? Should the commercial infrastructure that powers the internet be kept secret? These are the sorts of questions raised by Mark Gibbs in his latest column in Network World. 'There is an alternate route available for nearly all services through Las Vegas or Northern California serving all facilities-based carriers in Los Angeles -- all interconnected at numerous L.A. and L.A.-area fiber-optic terminals supporting both metro and long-distance cable.' Given that the internet thrives on open networks, it's hard to imagine keeping them a secret. At best, we must be prepared to deal with the worst."
There is already a tag which our software recognizes as indicating a typo in an article. It's 'typo'. This is in the FAQ. If you want to get the attention of the editor on duty, use the 'typo' tag.
As a consultant I routinely receive requests for Disaster Recovery work for organizations ranging in size from a few hundred to a few hundred thousand. Depending on the alloted budget we work our way down the hardware.
1. Redundant Network Connections
2. Highly available Services (Applicaiton Clusters)
3. Fail over - Off site if needed (Local, Metro, then off-site)
4. Power backup & Isolation (Generators good for 48 hours at least if not more, plus filtration systems that will withstand a localized EMP)
5. Testing - Smoking hole scenarios. (ie: where did NY, Chicogo, Washington, just go?)
I am not at liberty to divulge my client list but I can say for certain that they are very interested in maintaining service availability even if their primary sites were hit directly by nuclear weapons. Services include all communications not just the internet. Arpanet was founded by the boys in green, they worry about these sorts of things.
It becomes a matter of balanceing function with cost, the old engineering addage does ring true here more than anywhere else:
Cheap, Fast, Reliable; pick any two!
Companies like Hugues, Teleglobe, and various governments of the G8 do what their budgets allow to facilitate redundancy, however since terrorism is a good political tool to motivate sales (along with natural disasters) then people in the consulting industry will be well met to help the organizations that make the internet redundant.
As for the power grid, Telcordia standards dictate that a carrier grade data center (if it's essential services) has to have some method of running even at a reduced capacity for extended periods of time. Thus there is a buffer provided for the local power company to get their systems working, that and most datacentres are close to large power supplies. This is the result of the original POTS standards. It's also the reason VOIP providers don't guarantee 911 service. The regulation and maintence costs on these datacenters is very high, which is how AT&T and Verizon justify charging an arm and a leg for your land line.
Then again, I've seen Tier 1 data-centerers undone by a fire-systems worker (plumber) dropping a wrench on the -48V bus-bar and having instantaneously weld to the A-Frame causing millions in damage and making an entire city core go quiet. Who needs terrorists when we have difficulty hitting 100% availability on our own, normally?