NSA to store yottabytes of surveillance data in Utah megarepository(crunchgear.com)
crunchgear.com
NSA to store yottabytes of surveillance data in Utah megarepository
http://www.crunchgear.com/2009/11/01/nsa-to-store-yottabytes-of-surveillance-data-in-utah-megarepository/
11 comments
Some quick math:
With 2TB drives, that means 500m drives spinning in the building, assuming no redundancy, fs overhead, OS data, etc. (but then again, who the fuck has this sort of production capacity? Definitely not WD or Seagate or any of the major players)
If you store then akin to a Sun X4540 server, you fit 12 drives per U of space. In a 50U rack, if you take away 2U for networking, that's 4812=576 drives per rack.
If we throw back in some redundancy, that's about 1m racks.
Each rack takes up about 10sf of space (~19"6', accommodating for clearances), so that's 10m square feet of datacenter space, not accounting for any sort of power, cooling, or logistics space.
Throw back in the extras and you're looking at roughly a mile square of datacenter space.
Wow.
So, we all understand this is bullshit journalism, right?
With 2TB drives, that means 500m drives spinning in the building, assuming no redundancy, fs overhead, OS data, etc. (but then again, who the fuck has this sort of production capacity? Definitely not WD or Seagate or any of the major players)
If you store then akin to a Sun X4540 server, you fit 12 drives per U of space. In a 50U rack, if you take away 2U for networking, that's 4812=576 drives per rack.
If we throw back in some redundancy, that's about 1m racks.
Each rack takes up about 10sf of space (~19"6', accommodating for clearances), so that's 10m square feet of datacenter space, not accounting for any sort of power, cooling, or logistics space.
Throw back in the extras and you're looking at roughly a mile square of datacenter space.
Wow.
So, we all understand this is bullshit journalism, right?
http://www.datacenterknowledge.com/archives/2009/10/30/more-...
"There’s been a lot of media coverage in the past week of the National Security Administration’s plans for a large new data center at Camp Williams in Utah (which we first noted back in July). The NSA is expected to invest more than $1 billion in the project, which is said to include more than 1 million square feet of space. "
"There’s been a lot of media coverage in the past week of the National Security Administration’s plans for a large new data center at Camp Williams in Utah (which we first noted back in July). The NSA is expected to invest more than $1 billion in the project, which is said to include more than 1 million square feet of space. "
From 2005 to 2007 to 2009, the largest available hard drive grew from 500 GB to 1 TB to 2 TB. This is the normal historical trend for hard drives[1].
Projecting that trend out a mere 10 years in the future, we anticipate 64 TB drives to be available, reducing requirements to 32k racks. 10 more years in the future and we have 2 PB drives, reducing requirements to a mere 1k racks.
The trend of course might slow before then, but the point still stands that if you're planning on a scale of a decade or two, those numbers stop seeming impossible and start seeming inevitable.
[1] http://en.wikipedia.org/wiki/History_of_hard_disk_drives#198...
Projecting that trend out a mere 10 years in the future, we anticipate 64 TB drives to be available, reducing requirements to 32k racks. 10 more years in the future and we have 2 PB drives, reducing requirements to a mere 1k racks.
The trend of course might slow before then, but the point still stands that if you're planning on a scale of a decade or two, those numbers stop seeming impossible and start seeming inevitable.
[1] http://en.wikipedia.org/wiki/History_of_hard_disk_drives#198...
To put it in context. A square mile has 27,878,400 square feet in it. The Empire State Building has 2,158,000 square feet of usable office space.
But this also assumes they're going to use HD as standard storage devices. They would probably be using a different technology. It's clearly not possible with current HD capacity.
They could also be building for the eventual capacity and not going to immediately support that much data. The comments on the article page makes this and other good points.
But this also assumes they're going to use HD as standard storage devices. They would probably be using a different technology. It's clearly not possible with current HD capacity.
They could also be building for the eventual capacity and not going to immediately support that much data. The comments on the article page makes this and other good points.
> They would probably be using a different technology.
not even the NSA can get a yottabyte of some new technology manufactured without it being all over the news.
not even the NSA can get a yottabyte of some new technology manufactured without it being all over the news.
The data doesn't have to be actively online. Tape technology is one option to reduce the amount of hardware and space needed.
and that process is fairly routine at most national labs and supercomputing facilities. Not highly available, but not sure these data need to be.
You're assuming they're using server units not mainframes. It would be interesting to see the difference in data center size between mainframes and racks.
edit: would someone like to explain why I'm being down-voted?
edit: would someone like to explain why I'm being down-voted?
The figure of merit here is really hard drives/square foot. Mainframes use the same hard drives as rack-mounted or blade servers, and it is unclear how much the computational substrate matters in the context of storage capacity per square or cubic foot.
Additionally, your wording seems to imply that you think a mainframe would be significantly more dense. While you may cut down on extra NICs, power supplies, cases, etc. by moving to mainframes, the high degree of redundancy in mainframes in general, and their storage subsystems in particular, may well cancel out this advantage as it pertains to density.
Additionally, your wording seems to imply that you think a mainframe would be significantly more dense. While you may cut down on extra NICs, power supplies, cases, etc. by moving to mainframes, the high degree of redundancy in mainframes in general, and their storage subsystems in particular, may well cancel out this advantage as it pertains to density.
10^12 TB is a lottabytes and, also, a yottabyte.
This reminds me of being in college not too long ago when a CS professor said "Keep this under your hat but the main server is going to have ONE TERABYTE in storage as of next month" and I got chills. 10 years ago it was a Big Deal that a major research institution had a machine with a whole terabyte attached to it -- not quite a supercomputer, but man, it had a TERABYTE, what else could you want?
These days teenagers routinely have access to terabytes, some fairly small companies deal with petabytes, and major government organizations have to go up a few orders of magnitude when they let their imaginations run wild about future needs.
This reminds me of being in college not too long ago when a CS professor said "Keep this under your hat but the main server is going to have ONE TERABYTE in storage as of next month" and I got chills. 10 years ago it was a Big Deal that a major research institution had a machine with a whole terabyte attached to it -- not quite a supercomputer, but man, it had a TERABYTE, what else could you want?
These days teenagers routinely have access to terabytes, some fairly small companies deal with petabytes, and major government organizations have to go up a few orders of magnitude when they let their imaginations run wild about future needs.
Am I the only one who actually thinks this is plausible? First, I don't think all the storage will be on hard drives constantly spinning and requiring electricity. Imagine huge tape spools like in the old days but with 10,000x more data on them and miles and miles long.
Also, every conversation, every satellite image at full res. There's a lot of data out there. Every website with every update. every database. every conversation, in every country.
They didn't say they'd be at a yottabyte next year, but no doubt, there will be that much information to store and the NSA doesn't work on startup timelines.
Also, every conversation, every satellite image at full res. There's a lot of data out there. Every website with every update. every database. every conversation, in every country.
They didn't say they'd be at a yottabyte next year, but no doubt, there will be that much information to store and the NSA doesn't work on startup timelines.
I'm actually glad the comments here are a debate on the technological aspects and there's little commentary here on whether this is right or wrong. This kind of commentary is exactly where HN should be and often is, I'd like to take the time to point out to everyone who regularly bemoans HN is turning into reddit: it clearly isn't!
I find the CG rhetoric to be aged and irrelevant. The paranoia is so prevalent they sound like two kids smoking up in a dorm room, they use no explanation and cite no information for the paranoia than 'its the man'. It's pathetic. It's a tech website and there's barely even a passing mention of the technological feasibility of this project, and what examination of the feasibility they have is very weak and based on poor assumptions.
The assumption that the NSA will have thousands of HDD permanently running is absurd. Quite literally 1 cubic meter of 50GB blu-ray discs easily stores ~2.75 petabytes of data at an expected lifespan of 10-100 years (this is based on pure physical dimensions alone, and not storage dimensions). A single 100 disk stack of 50GB BR discs should store ~4.8TB of data. This makes the mass storage solely dependent upon write capability rather than size requirements.
I find the CG rhetoric to be aged and irrelevant. The paranoia is so prevalent they sound like two kids smoking up in a dorm room, they use no explanation and cite no information for the paranoia than 'its the man'. It's pathetic. It's a tech website and there's barely even a passing mention of the technological feasibility of this project, and what examination of the feasibility they have is very weak and based on poor assumptions.
The assumption that the NSA will have thousands of HDD permanently running is absurd. Quite literally 1 cubic meter of 50GB blu-ray discs easily stores ~2.75 petabytes of data at an expected lifespan of 10-100 years (this is based on pure physical dimensions alone, and not storage dimensions). A single 100 disk stack of 50GB BR discs should store ~4.8TB of data. This makes the mass storage solely dependent upon write capability rather than size requirements.
My math shows they can store 126,419,753,08,641 hours of uncompressed 640x480 video. This is based on 1 hour of video taking about 79GB of data.
Using basic compression can take the number of hours stored much higher.
Using basic compression can take the number of hours stored much higher.
Which is only about 1 billion years of video. Presuming they wanted to video every person in the United States (let alone abroad) you'd fill up that many hours in a little less than 4 years.
Now, you do get a nice multiplier from compression (10-100x, depending on the video and the quality) but that only brings it up to 100 billion years of video, max. What if you wanted to store all the video ever recorded? Bear in mind that the amount of video being recorded is currently in the exponential increase growth phase.
Now, you do get a nice multiplier from compression (10-100x, depending on the video and the quality) but that only brings it up to 100 billion years of video, max. What if you wanted to store all the video ever recorded? Bear in mind that the amount of video being recorded is currently in the exponential increase growth phase.
Depending on your goal, one doesn't need to store every single frame of a video in order to capture its plot.
Right, definitely not, but this is leaving aside all the other kinds of data you might want to save, and there's a lot of that too. I'm just pointing out that the quantity of hours we're talking about is not nearly as absurd as you'd think at first.
But clearly they want to track everyone and everything everywhere. Scares the crap out of me.
There is talk of the NSA shutting down altogether or being rolled into another agency,
Seriously?
Seriously?
Nah, they're thinking of the CIA.
...it approached the point where there was no substantive difference between the Library of Congress and the Central Intelligence Agency. Fortuitously, this happened just as the government was falling apart anyway. So they merged and kicked out a big fat stock offering.
...it approached the point where there was no substantive difference between the Library of Congress and the Central Intelligence Agency. Fortuitously, this happened just as the government was falling apart anyway. So they merged and kicked out a big fat stock offering.
quote from the article's source, itself quoting a non-online source: "a projection that sensor data volume could potentially increase to the level of Yottabytes (1024 Bytes) by 2015"
vapor.
vapor.
source: http://www.metafilter.com/86293/Yotta-vote-against-this
question: via moore's law, if it can hold up, what year can i hope to get a yottabyte on my netbook? thx
question: via moore's law, if it can hold up, what year can i hope to get a yottabyte on my netbook? thx
Assuming your netbook has a 32 GB SSD, and that drive capacity doubles every two years, 2099. Assuming that SSD prices drop dramatically and you have a 512 GB SSD (still this year), 2091.
In other words: not for a long, long time.
The real question is: how could you possibly generate a yottabyte of data?
In other words: not for a long, long time.
The real question is: how could you possibly generate a yottabyte of data?
All in one place?
This seems like a hugely un-necessary thing to do; I imagine the next headline on the news will be "NSA rescued from 2 square mile apartment filled with 8000 cats and pizza boxes to the ceiling".
I honestly can't think of any technology that is stable enough to support this magnitude of long term storage.
Dealing with the amount of bit rot alone leads me to boggle at the notion.
I honestly can't think of any technology that is stable enough to support this magnitude of long term storage.
Dealing with the amount of bit rot alone leads me to boggle at the notion.
"As the sensors associated with the various surveillance missions improve, the data volumes are increasing with a projection that sensor data volume could potentially increase to the level of Yottabytes (10^24 Bytes) by 2015."
Note the qualifiers and the future time date; also, this is talking about raw sensor data. It is not uncommon for raw data to be reduced by a few orders of magnitude via aggregation, compression, and simply discarding irrelevant sensor readings before hitting permanent storage (e.g. look at the LHC).