Alright, here’s the setup: I’m the ruler of an new Orwellian world dictatorship, and I want to start up your standard privacy-invading ID system.
Basically, all citizens (starting with the major 1st world countries, and working down from there) would have their photograph, fingerprints, DNA, and dental records recorded, and plugged into a searchable database. (Along with DOB, height, eye color, etc.—I won’t worry about getting the Yanomami indians into The System for awhile.) They’d probably be issued with a personal ID card, too, but I don’t imagine that’d have to have DNA and dental records stored on it.
Anyway, my question is…how much would it cost to gather all that information, per person? And how much cost (and labor) would it take to set up the database?
This is, of course, leaving aside all moral, ethical, etc. considerations. I’m just interested in material expense.
The way I see it, this task should be broken down into two distinct phases.
Gather the data (or at least insure it exists for all citizens)
Centralize it in a searchable database.
3.*
According to this cite, the last US Census cost $16 per person. Let’s increase that to $100 per person, for the following added costs:
[ul][li]Everyone actually has to physically be fingerprinted/photographed, etc. For the census, they just have to talk to someone in each household.[/li][li]Jackbooted thugs to accompany the identifier around to each house.[/li][li]*[/ul][/li]In the third world, there are fewer people who would already have things like dental records, but this is balanced by the decreased labor costs. This provides an estimate of $660 billion.
Of course, there may be other factors. If you are sufficiently feared, you could just force all citizens to fill out a form, fingerprint themselves, and mail it in. You might, like the SDMB, be able to fill the jackbooted thug positions on a volunteer basis.
For an upper bound on sorting and organizing the data, Google’s market cap is $125 billion. I bet they could do it in a few weeks.
There is no Opal. There has never been an Opal. Saying “Hi” to an unperson is crimethink.
All I can comment on is the backend database itself. The technology to run it is already commercially available (Oracle for example) and the hardware like massively parallel Unix servers are readily available and up to the task.
The stuff you list that you want in the database is just standard info and the software architecture needed is already there. Many websites have databases that do the same thing but they are just smaller. The very tricky part is the DNA sequence. That throws the whole thing into another dimension and I am not sure that can be done well at the moment. Where are all of these sequences coming from and how to they get into the database?
The software might cost $10 million to buy and at least that much to license it every year after that. It could really be anything above that though depending on the agreement. Modification and data architecture might cost $10 million. If you think these rates seem high, I used to work for a retail company that blew $24 million trying to replace just one of their database systems on their own.
The hardware might cost $10 million up front and a few data centers to house them. That is several million dollars a year for the buildings and more for the employees.
Costs tend to spiral out of control once you get past a certain size because it starts to become a whole entity in its own right that needs everything from security to janitors to systems analysts.
Security would be a major concern here so you need top notch everything including facilities and procedures and security people that test it for weaknesses all the time.
That is just the database itself. Would it be accessible through the internet or by some other means?
I guesstimate $100 million a year just to run the database itself.
As I’d envisioned, it’d mostly be a government-use system. Like if the police recovered fingerprints or DNA from a crime scene, or if someone discovered a badly decomposed corpse in the woods, they could automatically run a search against the database to show who came up. (I understand that similar systems for searching for fingerprints and DNA already exist, albeit in a slightly more limited scale.)
I didn’t imagine a lot of use for civilian users…but, then again, I perhaps I could see subscriptons/licences to use the system to people or companies. (Like for background check places, or for an ID reader for stores—the latter wouldn’t have to give the user more than the photo and age of the subject, I suppose.)
So roughly $100 mil a year, for the database itself? That sounds affordable—I was afraid it was going to upwards of tens of billions a year!
Well, the human genome contains about 3 billion base pairs, and one byte of computer storage can store 4 base pairs. That means that given six billion people, it’d take 9e18 bytes to store all the DNA sequences. From Froogle, $80 seems about reasonable for a 200 gigabyte hard drive, so we’d need 45 million of those, for a total cost of $3.6 billion. Of course, that’s an up-front cost to set it up. The total cost won’t repeat, but hard drives wear out. If we assume that each hard drive lasts on average for 10 years, then each year, we’ll have to replace $360 million worth of hard drives. That’s over twelve thousand hard drives replaced per day: Assuming that it takes 15 minutes for a technician to replace a hard drive, and the technicians work 8 hours a day, we’d need 385 technicians maintaining the system. Assuming a salary of $50,000 for technicians, that’s another $19 million per year. And, of course, we’d also have a similar proportion of technician work to set the system up initially. The bookkeeping of keeping track of which drives fail, and transfering the data to and from backups, I’m assuming is all automated (think a giant RAID array), and the development of those automated systems would be a negligible one-time cost.
So my total figure for the operating budget of the DNA database is $379 million per year, plus a start-up cost of ten times that. I suspect that the cost of the sequencing itself would be much more significant, but I have less to go by on that, and I don’t think the infrastructure is really in place yet for mass sequencing of large numbers of individuals, so I won’t comment further on that aspect.
Interesting…but do we need to record that much genetic information* just for identification purposes? I was under the impression that you didn’t have to record every single nucleotide when doing DNA fingerprinting?
*Not that I wouldn’t like to have that all on hand. God, how I’d love to have that much on file…
Perhaps I could restrict the more complete genome recording to, say, all new live births? (Of course, I’ll just have to spend the same amount of money on the DNA database, eventually, but at least I’d have more time to prepare and ramp up to it.)
God, I would love to have one of those. Not for our new Ranchoth overlords, but for the decomposed bodies we find in the woods… and the bay… and in abandoned houses… and in their own houses where their children never visit them, but we have to PROVE it’s them, and they’ve never been fingerprinted, and they haven’t been to a dentist since before they lost all their teeth…
The only searchable DNA database we have right now is on rape and homicide accused and convicted prisoners. It’s a very small percentage of the total population. The only reason we get hits is of recidivism.
There is a central fingerprint database, thank God (actually there’s three - law enforcement, military, and civilian - have to beg all of them for separate searches - I’m in the civilian one because I was printed as a condition of starting a job once, but not in the law enforcement or military databases). But the vast majority of people have never been printed, so there goes your hope of ID. And fingerprint databases take a long time to search; the automated system has to compare a lot of visual points to try to ID, and it’s imperfect.
No, we could do RFLPs and store a lot less information, but the more people you have in the database, the more likely you are to have two sets of RFLPs coincide. It’s like the Bertillon measurements that preceded fingerprints.
It’d be quite easy to get all the new live births that occur in hospital. Docs/nurses collect cord blood from every infant anyway. All you have to do is enforce it. The problem is enforcing all the ones that occur at home.
I’m not able to make cost estimates, but it strikes me the emphasis in these responses has been on the database, and from the little I know of police work I will just bet it would be a whole lot more expensive than $100 a person to arrange compliance with the testing of the entire population. The Census is benign. Go ahead and catch all your rebels, your Montana we don’t trust the govt activists, your Alabama we hate the feds more than black hates white traditionalists, your home schoolers and home birthers who’d rather you not know how many kids they have, your Mexican immigrants, your spy families, your poor people who move a lot, your homeless people, and your criminals. I think compliance with 50% of the population would be easy and maybe you could do it for a few hundreds per person; people like me, who have to live their lives licensed and regulated and OK’d by the govt, would be the first to crumble and comply. But between fifty and seventy percent would take a whole lot of work, seventy to ninety much more, ninety to ninety-nine much more than that, and you’d never catch the last percent. I think eventually most of the revenues of your police state (including salaries and benefits for your millions of goons) would be devoted to trying to insure compliance. And then how do you build up your enormous army for pre-emptive self-defense against non-conquered nations?
Very good points…I’ll have to remember to hire you on as an advisor once I’m emperor.
Of course, depending on the evil government, I’d guess that getting to a few of those “problem groups” might not be so much of a problem—with, say incentives for ID volunteers (government cheese; tax credits; free Big Brother Brew malt liquor Now with Neutersol™!, etc.); mandatory ID processing upon arrest; punitive atomic liquidation of rebel infestations, etc.
But still, good advice on not bankrupting the empire for the sake of an unattainable quota. I’ll keep that in mind—unlike some of the others doofuses in the king line. (Stalin and Mao, I’m looking at you, here!)
Yeah, if you only want near-universal compliance, it gets cheaper.
Remember if you give me a room with running water and an endless supply of sharp blades to dismember corpses, I work for free.
No, wait, am I a henchman again? Gonna be clash of the overlords time soon here. Maybe I better get a ringside seat. Or maybe I better get outa here leaving smokin’ footprints.
The British government is planning to do almost exactly what you propose: a national database of biometric information, with each citizen carrying a card with a portion of that information on it. However, they’ve elminated the census costs by integrating the card distribution with applying for passports. Even so, the database is estimated to cost at least £10 billion, according to relatively conservative estimates by the LSE. Considering that Britain is a relatively small country, I think this system would be prohibitvely expensive.
