System Created to Give Users Control Over Their Private Data

[Guest_Jim_*]

Unless you are particularly ruthless at culling and controlling your devices, there is probably a host of applications and apps tracking you and recording the data, possibly to upload to a server. Where that information is stored, whether it is being kept secure, and if it is being shared with other systems is something the user might never know. To at least demonstrate a different infrastructure that puts the user in greater control, researchers at MIT have developed Sieve.

With Sieve, the user uploads their information to the cloud, in an encrypted form, and then whenever an application wants to access the data it needs to get a secret key from the user to decrypt it. To make this work Sieve uses attribute-based encryption and key homomorphism, which are two new cryptographic techniques. Attribute-based encryption applies labels, or attributes, to different pieces of data so that secret keys can be generated for each label, and not the larger dataset. If an app is only asking for your address, the key will decrypt that information, but not your date of birth. Key homomorphism is necessary for scenarios when the user might want to revoke access to the data. This is done be changing the secret key for the data, and key homomorphism allows the data to be re-encrypted without first decrypting it.

Sieve will need app developers to support it before it can be commercially deployed, but in some cases the developers may benefit from this support. This benefit would be from the apps having access to data from other devices, also stored in Sieve.

Source: MIT

Back to original news post

[Guest Guest]

Hmmm

 

I find the notion that the solution to localized data integrity is to introduce yet another external independent variable to be illogical at best. Considering that the likes of MIT is the source of this strategy, it is even more suspect.