Recent breakthroughs in data encryption allow you to do calculations on encrypted data, without knowing the encryption key.

This is called Homomorphic Encryption. You can do operations such as:

MagicAdd(Encrypted(num1), Encrypted(num2)) = Encrypted(num1 + num2)
MagicMultiply(Encrypted(num1), Encrypted(num2)) = Encrypted(num1 * num2)
# Without knowing the encryption key!!

This can unlock tremendous business value e.g.
• Processing secure data in an insecure cloud
o Medical, genetic, financial, legal, voting, …
• Use a trained neural network to do predictions on encrypted data e.g. image recognition on X-Ray images

The goal of this project is to work with our security experts:
• Experiment with homomorphic encryption and/or multiparty secure computation
o Student can choose to work more theoretical or more practical
• Build a demo which processes encrypted input data and produces encrypted output data (e.g. a mathematical calculation or a machine learning prediction)
• Integrate this demo in the Western Digital ActiveScale storage system such that encrypted data is written into the data storage system, and some encrypted calculation is returned

Technology
• C/C++ programming (Python and Java also possible) to make a cool demo
• Finding open source libraries to use e.g. https://github.com/shaih/HElib or https://github.com/n1analytics/python-paillier or https://github.com/Microsoft/SEAL
• Apache Kafka API for getting notifications about new data to be processed

More information & very informative slides

• https://www.cosic.esat.kuleuven.be/events/industryday2018/slides/Paillier.pdf
• https://www.cosic.esat.kuleuven.be/events/industryday2018/slides/Smart.pdf
• https://venturebeat.com/2018/12/03/intel-open-sources-he-transformer-a-tool-that-allows-ai-models-to-operate-on-encrypted-data/