Posted Wednesday 20th May 2009 22:48 GMT
"And it will have its own dedicated Single Point of Failure with two links to the power grid for redundancy."
If you have 1 sub station, it doesn't matter HOW many links you have past that, if the sub station shits-itself-and-dies, then its a moot point. It would be better to have two small sub stations, on opposite sides of the building, each with 2 links to the outside world. THAT would be proper redundancy. I never said it was cheap, but it would be close to failsafe as you could get.
So really, to take out an entire data centre, you throw a firebomb into the sub station, and watch the sparks.....
Mines the one with the molotov cocktail in the pocket.....
Posted Thursday 21st May 2009 15:33 GMT
Andrew, It isn't clear [to me] if the article is saying the sub-station had links to the grid, or if the facility had the separate links?
of-course if the substation has the links and then dies, then yes, single point of failure, and we just need to sit and wait here for that story where an entire cloud is, er, vaporised!
DHD
Back to the article