Homomorphic encryption has been around as a concept in academia, but it’s only now that it's starting to be used in business. Two things have made that possible: this form of encryption has gotten fast enough, and it has become scalable.
The global homomorphic encryption market, valued at US$117.8 million in 2016, is expected to grow to $ US$268.3 million at a 7.55% CAGR by 2027.
Homomorphic encryption is not an easy concept to understand, even for data analysts.
Here’s a quick primer…
Performing analytics on encrypted data is not possible for obvious reasons. The standard encryption method involves encrypting the data at one end and decrypting it at the other using a “key”. This generally means that only the key holder may decipher the message. Because the data is under lock and key, it is not visible to anybody except the person who holds the key. Because of homomorphic encryption, though, such encrypted data can now be analyzed and worked with without compromising the integrity of the encryption.
Some explanations of homomorphic encryption:
“A form of encryption which allows specific types of computations to be carried out on ciphertext and generate an encrypted result which, when decrypted, matches the result of operations performed on the plaintext.”
Here’s another:
Homomorphic encryption is a secure encryption scheme that enables the computation of functions on ciphertext to produce an encrypted output. It aids in managing the data to have authorized access without compromising the data.”
So there it is. At the risk of repetition, homomorphic encryption enables the computation of functions on encrypted data without decrypting it, thereby preserving data security.
This also means the holder of the decryption key does not necessarily have to be in the same room, figuratively speaking, as the person doing the analysis.
The idea behind homomorphic encryption can be traced back to the early ‘80s, when MIT professor Shafi Goldwasser proposed a new hypothesis in cryptography: that one could prove something was true without disclosing anything about that “something”.
This would lay the groundwork for much of modern-day cryptography and earn the professor the Turing Award, considered the “Nobel Prize” of encryption.
Data security is one of the few hurdles to Cloud computing. One common refrain among organizations is that they are not adopting the cloud due to security concerns.
That’s a fact because processes like mining and data analysis in the Cloud cannot occur if the data is encrypted. And organizations such as hospitals or medical research facilities that are custodians of sensitive data, such as patient confidential information, cannot use such data to perform any data analytics. Homomorphic encryption will now help solve this fundamental problem.
Possible Use of Homomorphic Encryption
This method can be used to analyze highly sensitive encrypted data, such as health and medical records stored in the cloud.
In this paper, published in the Journal of Biomedical Informatics, the team examined potential application scenarios for homomorphic encryption to ensure the privacy of sensitive medical data.
Planning your next-gen data security roadmap? Find out where homomorphic encryption fits in >>>>>> Learn more
It eventually demonstrated a working implementation of a cloud service for performing private predictive analysis tasks on encrypted health data using homomorphic encryption. The cloud service makes predictions while handling only encrypted data, learning nothing about the confidential medical data submitted.
This means analysts can accurately predict the likelihood of contracting a disease using only an encrypted medical record.
Even in some of our day-to-day businesses, homomorphic encryption can be put to good use, because data privacy has become a big issue of late. New data privacy laws, such as the European GDPR and the American CCPA, have been enacted to ensure data privacy.
There are services out there, like encrypted chat or email services, that assure complete privacy, and by their very nature, lose out on the analysis part. Homomorphic encryption will now allow that.
