I hate how this shit is even in computers at all, much less networked ones. Can we not just revert to using paper and folders and filing cabinets for stuff like this until hacking computers isn't so trivial?
Are you suggesting that one of the largest HR depts. in the world revert to paper & filing cabinets? Can you imagine the massive cost of that? The efficiency loss? It would be staggering.
Governmental record keeping of this sort is literally one of the first uses for computers.
This shouldn't have been stored on an unclassified, publicly accessible computer. This is a prime example of how the aggregation of unclassified information can be used to derive classified information.
Example... oh, there's lots of people with TS clearances in this little town in the midwest. Maybe there's something to that. Let's check it out.
Apparently during ww2 the editors of a SF mag worked out something nuclear related was going on at los Alamos by looking at the cluster of subscriptions
I really do not think the efficiency loss will be large.
One centralized location that you can phone in with some smart authentication system, human workers for the rest of it. Getting cleared already takes a while (weeks or longer) so what's the huge problem with keeping the classified stuff actually safe?
At this point I don't feel especially confident that any of this stuff is going to stay secure given the prevalence of leakers and hacks. The choice is store on a computer and accept that it is going to be leaked at least occasionally or do not store it on a computer.
There were 21.5 million records stolen. If each file averages 5 mm in depth thats over 100 km of depth you need to have (you can do similar calculations for height/width).
If the OPM computers currently average 1 query per second (a number I'm pulling out of thin air, but one I'd guess was conservative). Even if you restrict that to 8X5 business hours access (an efficiency loss in itself) you have 2400 phone calls to answer and then traverse of the gigantic document store we created above.
I think this is one of those cases where people have very bad intuitive understanding of the scale of the data and what dealing with this kind of data looks like in a physical form.
