Algorithms and Homomorphic Encryption will protect you from Big Data breaches
In theory, Big Data is a brilliant idea. In practice, I worry about Big Breaches by Big Government and Big Corporations. If we are going to hand over power we need to see it handled responsibly. I don’t mind someone looking at my gene code as long as they’re using it to extend my life, rather than let someone else ruin it. In 2016, news coverage of Big Data seemed to go hand in hand with mentions of state sponsored hacking and ID theft.
Before people warm to the idea of Big Data, they need to be able to associate it with positive interventions, such as the pre-emptive treatment for diseases. Not negative interventions, such as invasive marketing campaigns or, worse, the emptying of our bank accounts. Yes, I’m talking to you, Yahoo!
The year 2017 will see concerted efforts to beef up security. One of the great hopes for the coming year will be new techniques to take security to new levels. Mathematics and cryptography in combination will bury our data deeper from the cyber criminals and state sponsored hackers. The only downside is this will make data mining a lot more processor intensive.
The two disciplines are currently being honed to protect the privacy of medical and genomic big data in hospitals. New homomorphic encryption schemes could make data secure enough to be put into the cloud and available to all the right people (trusted scientists) while still being excluded from the hackers.
I don’t know about you, but before I put my genomes on the line I want to know exactly how Homomorphic Encryption will protect my privacy. This is a subject that needs to be explained to the public as it’s their data. Telling them that there’s a new form of encryption, that allows computations to be carried out on Ciphertext, isn’t going to clarify things and win their confidence.
At the risk of over simplification, this new form of security is akin to keyhole surgery, without having to make an incision to get access to the organ being operated on. First data is converted by an algorithm, known as a Cipher, into ‘Ciphertext’. Though it is rendered unreadable to the outsider, this encrypted information can still be analysed by a Homomorphic system.
The beauty of Homomorphic encryption is that it allows the Big Data analyst to chain together different services without ever exposing the data to each of those services. So a pharmaceutical company – having completed its researches into genetic variation and susceptibility to different types of disease – could perform calculations on where to develop drugs and perform clinical trials in one go. It would not need to open the data up – and expose it to the risk of theft – each time it needed to perform calculations on, say, tax, currency exchange rates and distribution, each of which calls for a difference service to be applied.
Homomorphic encryption schemes are designed to be malleable enough for cloud computing. The main challenge that pioneers face is that it needs a lot of computing power to process information that has been deliberately rendered into a form that’s designed to be inaccessible. So the carbon footprint of the data centre industry is going to expand dramatically if Homomorphic Encryption takes off. Think of all the millions of central processing units in each of the hundreds of thousands of data centres across the globe. They will all be working overtime, meaning that more cooling systems will be needed to chill them, burning ever more fossil fuels to power them.
However, there doesn’t seem to be any alternative. In December 2016 encryption Intelligent Voice claimed that only 25 percent of people trust their cloud providers not to misuse their data. The basis of this claim was a survey of 1,500 American consumers who were asked how they felt about handing their data over to major cloud companies such as Google, Apple and Amazon.
“People want a world where their data is safe from abuse by the people they trust to hold it,” says Nigel Cannings, CTO of Intelligent Voice. New techniques like homomorphic encryption could encrypt their data before it reaches the cloud. However, since that would prevent Big Cloud from advertising against it, none of the infrastructure is being put in place.
The danger’s not over yet!