Although advances in data analytics have enabled businesses to acquire expanded insight into large structured and unstructured datasets, these advances have limited privacy and misappropriation risks.

Having greater control over the life cycle of data and confidentiality agreements has alleviated these risks, but outsourcing sensitive or regulated data processing components to third parties is still broadly viewed as fraught with risk.

Any third party, including competitors, could share all sensitive data or data processes and algorithms, subject to the provider’s controls. However, it would open unimagined avenues of enterprise collaboration, integration, and specialisation.

Homomorphic encryption voids this significant gap. As commercial viability is still a challenge, compelling use cases are emerging. In the coming years, any organisation that tends to become a center of excellence in big data analytics will be left with no choice but to embrace homomorphic encryption.

Encryption and its Limitations

Encryption is a digital locker where information is secured until locked inside. Plaintext data is converted to ciphertext, applying a sufficiently complicated algorithm to make the data unreadable without a decryption key. Once analysis, compliance, or any other use case requires encrypted data, it must be converted back to plaintext that can compromise security.

It addresses the core weakness by allowing the analysis of data in its ciphertext form. An early homomorphic encryption innovator, Craig Gentry, described the process by manipulating a locked box’s contents through gloves that are accessed through ports outside of the box.

It’s difficult for a third party to access locked content or what another party is working on. The box is returned to the controller once the processor has completed the assigned task, and custody is intact.

Gentry’s dissertation made homomorphic encryption achievable, but there’s a significant barrier, which is computational overhead. Processing ciphertext makes a lot of overhead as the calculations are done bit by bit. IBM claims that it has improved processing overhead and now runs 75 times faster than before. A broader range of alternative schemes has notably improved processing speeds.

#big data #latest news #data-science