The best health-care advice you can get is: “Eat right, sleep right, exercise, wash your hands, do Yoga, and don’t watch the news.” We have plenty of data to validate such advice, and it works. Alas, every so often, we get sick, fall, get hurt. And then we expect a powerful health-care system to make us whole again.
Now imagine we diverted all the investment in doctors and hospitals to make roads less slippery and to send the food police to make us all eat broccoli. Fewer of us would require hospitalization, granted, but what about those who get sick or wounded nonetheless?
Remarkably, in cyber security this ridiculous strategy is the norm. Here, the “Stay Healthy!” equivalent is “Don’t Get Hacked!” which, too, is very good advice. But what if you do get hacked, what then?
Publicly available databases (the kind merchants and banks run) must keep their data ready, readable, and searchable. Yes, we surround our data with cyber fences, but we are also surrounded by diabolic cyber intelligence.
The good guys may win 99 times, but if they lose once, they lose their shirt. The most protected databases in the world have been embarrassed by humiliating penetrations.
We need heavy strategic emphasis on bouncing back, on minimizing, even neutralizing the damage of a broken cyber fence. Some principles: (i) data-at-rest-encryption; (ii) high-resolution, fast-rewind backup technology, and (iii) data fingerprinting. Combined, these three tools will make hacking uneconomical for the hacker and drive those smart computer mavens to an honest job.
End-to-end encryption is a great buzzword, but the big damage from hacking is due to theft of data at rest. It must be encrypted. When pulled for human review, decrypt it; when returned to storage, re-encrypt. But what about statistical reports? If the data is encrypted how can one find patterns therein? A modern branch of cryptography offers effective solutions to this challenge. Also, new super-fast ciphers reduce any delay in data retrieval to alleviate any inconvenience.
Data backup has been a routine protocol, but backup technology has not kept up. Archiving an image of a database every so often is laborious. Event logs are more efficient but rarely applied at sufficient resolution.
Recent technology allows a database to secretly mark the identity data of its customers. Dubbed “fingerprinting data,” the idea is to effect a subtle distinction between the private data held by its owner and the same data stored in the service database.
When the database is hacked, the thief would use the subtly marked data to attempt to steal the victim’s identity, but unbeknownst to him the marked data will tell the merchant or the bank that the data presented came from a hacked database. The attacker then turns prey. Once hackers realize they can’t profit from their spoils, they will lose their incentive to attack financial databases.
These powerful recovery strategies should be combined with new, efficient ways to spot penetrations, which often remain hidden for months. The new digital-money technology may prove handy. Much as “Follow the Money” is the prime fraud-detection technique in the real world, so it may be for cyberspace, once we implement fast, frictionless, high-resolution digital payment for every cyber service.
Effective cyber recovery requires data refreshment, which calls for shying away from biological and other immutable identifiers, relying instead on fully randomized, expiry data. Inconvenient, yes, but we give up convenience to buy security. None of the great promises of cyberspace can be realized if we lose the security battle.
—By Gideon Samid, gideon@bitmint.com