jhungary
MILITARY PROFESSIONAL
- Joined
- Oct 24, 2012
- Messages
- 19,295
- Reaction score
- 387
- Country
- Location
Obviously an insane amount of data has to be processed through supercomputers and humans can only look at the pieces flagged by it. We don't expect anything different from our intelligence agencies. You say that billions are required for this arrangement but I personally think it can be done on a much smaller budget than what the US spent. The reason being that cost of intelligent labor like Software Engineers etc in US is so much more than countries like China and Pakistan. Cost of building supercomputers would be a little less as well.
And considering the dire need of a capability like this Pakistan, I don't think spending $500-600 million or even a few billions is too much to ask for. We already spend tens of billions on defense every year. This wouldn't matter much.
And just for the record, the world fastest supercomputer, Tianhe-2 , cost around $390 million to build. Obviously no one in Pakistan would be looking to surpass that amount or speed.
Well, this kind of activities need more than just a supercomputer....
In the intelligence cycle, you need a way to gather data, a way to process data, a way to dissimilate information, a way to protect those information and a way to react to the information. All this 5 stage form the core of intelligence activities.
Supercomputer is only a part of ways to process the raw data, and then you will need more than a single computer to decipher all the data you need as you are facing 3 different set of raw data. Text (E-mail, Webpage), Voice (Phone Call and Video Call) and Image (Video Call and surveillance photo). You need to have 3 separate supercomputer to process all these information.
Then you also need to look at your ways to collect data. Equipment wise, you will need a filter box physically attach to any cell tower or telephone exchange to basically filter the call to your agency, then you will need a communication network which mirror your local telephone network to basically put transport those volume of call to your agency.
But what if the target you are tracking did not just use landline and/or cell phone? Even in the US, not all area are covered by Cell tower and Landline, which mean microwave equipment and satellite equipment would need to monitor CB/Radio Band wave and Satellite Communication. And also one-way communication (Such as broadcast or number station) would require good old fashion Man power to monitor. Surveillance wise, you will need CCTV in street corner where you did not had them already, and where you had CCTV already, then you will need add them to your surveillance network.
Then Data Processing, beside the different Supercomputer you will need to process those data, You will also need a data farm where you can store all the data (both relevant and irrelevant), communication between data farm, people to work at the data farm. And security protocol within the Data Farm. Depending on network security solution, each data farm would need to be turn into a different level of SCIF, which would cost different set of money for different level of SCIF.
After you have your data processed, then you will need to find a way to independently verify the data for its authenticity. Not anyone saying "Bomb" is automatically a threat you know? But those keyword will be filtered so they will need to see who is saying it and what is saying it for. That mean a team to verify the information and see if they were genuine threat.
Afterward, you will need to protect all the step you did above, otherwise your target may be able to penetrate your network and hence negate your whole operation. You need to do that by a way to implement OPSEC, INFOSEC and COMSEC. Depending on the level of security you want, it will again cost you a lot of money.
Depending on the level of information you want, the coverage you want and the level of security you want, the money factor is quite different. For a full spectrum one like the NSA were using, although the figure is classified, the estimation to establish a PRISM like system cost upward to 5 billions, and the yearly operating cost (Before operational support) are 20-25 millions a year.