Homomorphic Encryption with Skyflow’s Avradip Mandal

Partially Redacted: Data, AI, Security, and Privacy

Imagine being able to perform any computational operation over any kind of data but do it while the data is fully encrypted. That is the promise of fully homomorphic encryption.

Fully homomorphic encryption was first theorized in the 1970s, but the first proposal for a plausible construction of a fully homomorphic encryption scheme didn’t arrive until 2009. We are now in the fourth-generation of fully homomorphic encryption and although performance is still a blocker for many applications, there’s been a series of major breakthroughs allowing real world application to take advantage of the approach.

Dr. Avradip Mandal received his PhD from the University of Luxembourg where his research focused on cryptography, in particular homomorphic encryption and theoretical symmetric key cryptography. He joins the show to describe what homomorphic encryption is, how it works, the history, and breakthroughs.
Topics:

• Can you explain what homomorphic encryption is and how it works?
• Where did this concept come from and what’s the history with it?
• What are some of the main advantages of using homomorphic encryption?
• In what types of situations is homomorphic encryption particularly useful?
• How does homomorphic encryption compare to other types of encryption, such as symmetric or asymmetric encryption?
• Why is homomorphic encryption so slow?
• Are there any challenges or obstacles that need to be overcome in order to fully realize the potential of homomorphic encryption?
• Are there real-world applications of homomorphic encryption?
• How has homomorphic encryption evolved over time?
• What impact do you see homomorphic encryption having on the future of data privacy and security?
• What advice do you have for organizations or individuals considering using homomorphic encryption?
• Do you think we’ll get to a point where fully homomorphic encryption is readily available and if so, when do you think that will happen?
• Besides homomorphic encryption, are there future looking privacy and security technologies that you are particularly excited about?