By Solomon GrundyPosted Wednesday 23rd April 2008 09:51 GMT
The solution is to just not store sensitive information on a computer. See, problems solved and the waste of resources that is IT is now under control.
By Simon WhitehousePosted Wednesday 23rd April 2008 14:56 GMT
"Functional encryption is designed to solve the hassle tied to traditional public-key encryption resulting from distributing and managing thousands or millions of private keys authorized people need to decrypt protected data. If 1,000 people in an organization need to securely share their public key with their co-workers, that requires close to one million separate exchanges."
You seem to be getting confused between symmetric and assymmetric encryptions. In a symmetric system you have a seperate secret key that you exchange with everybody. Total number of keys needed is n(n+1)/2 which does get very near to a million for 1000 employees.
In an assymmetric system you have one public key that is available to everybody. There is no need to secure your public key, as it's a public key: there is a clue in the name after all. In an assymmetric system you only need n keys.
By KanhefPosted Wednesday 23rd April 2008 15:00 GMT
by some financial institutions. Any account details are sent in a pdf, password-protected with the account holder's SSN. Not the same as encryption, but it's a similar principle.
By Dan GoodinPosted Wednesday 23rd April 2008 15:35 GMT
Ever heard of key signing parties? Even with asymmetrical schemes, there is a need to securely exchange public keys. If a bad guy fools me into using the wrong public key to encrypt a message, the entire system fails.
So yes, an organization of 1,000 people still need to figure out a way to securely distributed their public key to each of their colleagues, and if you do the math, that's very close to 1 million exchanges.
By Chris CPosted Wednesday 23rd April 2008 19:40 GMT
Maybe I'm just being dense here, but I always thought the purpose of encryption was to make sure people viewing the data couldn't decipher it. If you use something as useless as a person's name or geographic location as a key (or even their social security number at this point, since that's not even close to secret/private anymore), then a brute-force attack to decrypt the data will take relatively little time. In other words, the encryption will be virtually useless.
Isn't this why we tell people to use complex passwords? Isn't this why we tell them not to use personal info in their passwords (own/spouse/child's name, birthdays, etc)?
And yet now we're telling people this is a good way to encrypt sensitive information?
By Chris iversonPosted Wednesday 23rd April 2008 22:58 GMT
Because its time to enforce a password reset: 26 character minimum, contain 5 numbers, and 4 odd characters. And remember to shoot anyone who leaves with their laptop.
By Daniel B.Posted Wednesday 23rd April 2008 23:47 GMT
True. Its kind of stupid using easily-guessed "keys" to encrypt sensitive stuff. Even if it is what I understand, names of different persons, there are finite number of combinations to be tried. On top of that, full names in English-speaking countries usually are name + 1 surname, which ends up having lots of collisions (just how many 'John Smith' can you find?) and reducing even more the probabilities. Which anyone would get by doing a simple SELECT * FROM employee.
@Dan Goodin: Thats what CA's were made for. You get *one* single Root CA, and do all the checks to verify the public key comes from the actual user. Place the signed public key on an LDAP, and the distribution problem goes away, isn't this what PKI is all about? ;)
Disclosure: My previous job involved Information Security, PKI, and other fun stuff.
By Anonymous CowardPosted Thursday 24th April 2008 07:16 GMT
"Mr George Clooney died this morning after the doctor treating him collapsed from exhaustion and another, drafted in from a different hospital due to staff shortages, was unable to access his medical records on account of administrative delays."
"It seems that the doctor had obtained the necessary paperwork, passed it to the Departmental Authoriser who, when he got back from lunch, passed it to the Senior Authoriser who, upon returning from his golfing holiday, approved it and passed it to an assistant, who faxed it to the offshore call centre that handles the hospital's IT admin. They queued, prioritised, processed and filed the request in accordance with their SLA, and then notified the doctor."
"However, by the time that all this had been completed, it seems that the unfortunate Mr Clooney had been dead for several days."
By Anonymous CowardPosted Thursday 24th April 2008 16:36 GMT
Anon, that's hilarious. That kind of administrative overhead is what keeps people from implementing this kind of fine-grained access with mechanisms available today (i.e. ACLs). No matter how strong your enforcement, you still need a person to sit there doling out credentials and assigning ACLs to the database records. This functional encryption scheme is just stronger enforcement but doesn't change the fundamental policy overhead.
Also, one point I made that wasn't in the article is that better auditing is important to resolving this issue. If staff is notified that a random user will be selected each month to have all their queries reviewed, then they will be more careful to avoid exceeding their authority. The staff that was sacked for looking up celebrity records were only found because there was a "lock" on those records. But this doesn't stop looking up boyfriends or bosses, etc. so there needs to be some random review process to protect those of us who aren't famous.
Comments on: Modern 'primitive' could ease the pain of encrypting massive amounts of data
Solutions #
By Solomon Grundy Posted Wednesday 23rd April 2008 09:51 GMT
Symetric and assymmetric encryption systems #
By Simon Whitehouse Posted Wednesday 23rd April 2008 14:56 GMT
Similar tech already used #
By Kanhef Posted Wednesday 23rd April 2008 15:00 GMT
@Simon Whitehouse #
By Dan Goodin Posted Wednesday 23rd April 2008 15:35 GMT
Am I being dense? #
By Chris C Posted Wednesday 23rd April 2008 19:40 GMT
good thing I read this #
By Chris iverson Posted Wednesday 23rd April 2008 22:58 GMT
Re: Am I being dense? #
By Daniel B. Posted Wednesday 23rd April 2008 23:47 GMT
Not ideal when time is of the essence #
By Anonymous Coward Posted Thursday 24th April 2008 07:16 GMT
George Clooney dead?? #
By Anonymous Coward Posted Thursday 24th April 2008 16:36 GMT