Does the NSA Need More Electricity? 324
An anonymous reader writes "The Baltimore sun (NSA watchers can't live without it) reports that the NSA worries about overloading the Baltimore energy grid if it were to install new computing capacity at it's Fort Meade HQ. This includes two multi million dollar supercomputers. Some systems are reportedly not getting the cooling they need. The temperature in NSA buildings is raised two degrees to conserve energy, according to the article. The NSA is Baltimore Gas and Electric`s (BGE) biggest customer the sun reports. Former NSA employees fear that a power outage at Fort Meade would have worse consequences than the 2000 "information overload" related outage. The NSA does apparently not have the backup power generation capacity to power the whole facility during power outages. Some point a finger at a new mall build in the area, but a BGE spokesman says the mall is "fairly easily accommodated". Some sources say the problem was identified in the late 90`s. But "keeping the lights on" wasn't a priority. A $4 million computer upgrade to the system that allocates power was postponed for budgetary reasons. (the NSA budged is estimated at $8 Billion) The article reports that the budget documents for listening posts around the world report similar infrastructural problems, in the budgets for 07 as well as previous years. It should be noted that the huge "groundbreaker" IT infrastructure upgrade program is reportedly over budget and late, but not yet fully operational."
Re:The NSA is a spy organization, but do we need i (Score:2, Informative)
People in power can attack/remove the persons and groups that might become a threath against their own power. This happens even if the threathing groups is working for the good of the nations and they will be elected trough the normalt democratic ways.
Democracy can be just as big threath against the president and his friends as anything else. Is more powe to the president then a good idea?
Why do people belive the power is allways working for them?
Re:Seems silly they don't have their own generator (Score:2, Informative)
Re:Back it up (Score:1, Informative)
Duplicating the power company is expensive.
Re:"We" do not want an efficient, effective govt. (Score:4, Informative)
Your post should be modded up... (Score:5, Informative)
I neglected to mention that my particular experience was in an embassy. Most of the administrative personnel are not American citizens, which means the USG is a bit leary about handing out cards to them. There were only a handful of American personnel running the Foreign Service Nationals (FSNs), and these Americans were very busy in other areas and weren't the type of personnel you ask to make small purchases.
Regarding GSA - we made big orders through them, but only once every quarter/six months or so (don't recall specifics). Shipping big pallets like that from the U.S. can get expensive, and between time it takes to make the purchase order, assemble the pallet, ship it to post, clear customs, etc. it can take months.
My experience will obviously vary from that of many federal employees.
The Truth about Classified Data (Score:3, Informative)
But there's a deeper truth, that I'd found while I was in the Air Force. 99.999999% of classified data is MIND-NUMBINGLY BORING. I did electronic warfare back then, and while it's useful to know the frequency ranges, etc of every type of radar you're likely to go up against, knowing that the BULL SHIT's radar's microfleems are, in fact, subradiate, is more than you really need to know, or even professionally care about.
(and yes, I did steal that line from Dilbert, just in case the radar data I remember from 25 years ago is still sensitive. . .you still have no need to know. . .)
But the bottom line on classified, is that you can probably get most of it from open source, the classified stuff is generally from "special" sources and we have a definite read on both the source of the data, and how good it really is. . .
Re:Toronto huh? (Score:3, Informative)
Did you look at the summer temperatures for both cities? They're about the same. The reason this works for Toronto is that Lake Ontario is pretty damned deep as well as big. I don't know if there are any lakes deep enough near Baltimore for this type of cooling to work.
Money. (Score:4, Informative)
And the reason they don't just build a small power plant is...?
Powerplants aren't cheap, and a well-designed UPS solution is a bit more complicated than just a spare power cord.
Using ballpark numbers, $21M/yr and $0.07/kWh gives about 40MW(e) load for our "UPS". Since you always buy one with room for growth, and since this is even more of a PITA to swap out than your average lead brick, call it 100MW(e) design. But how do we make it uninterruptable? Remember, all powerplants have maintenance downtime, even without accidents. You ideally want three plants, any one of which can handle the full design load; this allows for one to be down for routine maintenance, one down for unexpected accident (say, a safety fault emergency shutdown), and one to keep the load going. You also want this as a "ready-swap" load, and a power plant does require a few minutes warm-up time (exact amount varying by type); so, you'll want to run them all regularly, and transfer surplus power to the grid, to offset (say) the Pentagon electric bill.
So, what kind of plant? Since we're doing three, I'd suggest using different types, so as not to put all the eggs in one basket, and further reduce the chance of a single-point of failure. I'll presume the Bolognium reactor from Area 51 is unavailable for this purpose. Wind and solar are too unreliable for this. Hydroelectric doesn't have a convenient enough water source. Geothermal is laughable in this location... although it might be a factor to consider for the Fort Meade Mk II location. Fossil plants (oil, coal, NatGas) have some environmental considerations, but not unmanageable; if necessary, designing the plant to liquify the entire stack output shouldn't more than double the cost. Nukes are compact, but REALLY don't like fast startups; a nuke also will make for an even bigger target for terror attacks... but of course, adding any power plant is going to paint an even bigger target on Fort Meade than there already is. Since it's really only adding another ring or two to the existing target, and since I'm not even halfway familiar enough to address such security considerations, I'll just ignore them. Some other Slashdotter can comment on that design aspect.
So, I'd pick a small nuke plant for the primary Meade power plant, with a liquid natural gas or oil-fired for one backup plant. IIR, it's not hard to convert between those two fossil fuel sources, which might be advisable if there are supply issues. The third plant might be either one; I'm not sure whether the higher simultaneous event/design failure risk of the nuke plants would be better than the additional fuel transport and security headaches for steady fossil fuel supplies. For argument, call it one nuke, one oil, and one gas, with the latter two designed with convertability in mind. I think the coal transport/storage would be the worst of the fossil fuels, so we won't use that at all.
So, we need three plants, each around 100MWe. My Googling suggests a pricetag of 100-200 M$ apiece for those; if you can find better numbers, feel free to note them. We need to arrange for steady suppies of fuel. We need to arrange for additional physical security. We need to find plant operators for all of them... every one of whom will probably need at least a Secret level clearance, to be confident a background check turns up anything nasty. We need to do some environmental work-up, since it's an urban area; "National Security" gets you only so far in Baltimore — although it might be enough to put a gag on the inevitable NIMBY idiots who'll turn out against anything.
So, we're probably talking half a billion dollars for the building of it, plus additional annual expenses including higher than average salaries for plant workers due to the need for a clearance. This isn't peanuts, even with the NSA budget. It's not a bad idea... but it's not the no-brainer it looks at first pass.
Re:tubes (Score:3, Informative)
At least Svetlana's RF tubes are relativly cheap.