Originally published in the French journal “LeMagIT”, this article delves into the innovative world of data encryption, with a special focus on the startup, Cosmian. The full French version can be accessed here.

In today’s digital age, data is the new gold. However, with its immense value comes the challenge of ensuring its security while also making it accessible for processing. This is where modern encryption techniques come into play, and several startups are at the forefront of this revolution.

One such startup making waves in this domain is Cosmian. Founded by Sandrine Murcia and Bruno Grieder, Cosmian is exploring the potential of homomorphic encryption, a technique that allows data to be processed without having to decrypt it first. This method promises a balance between data security and usability, a combination that many businesses are eagerly seeking.

Bruno Grieder, Cosmian’s technical director, highlights the allure of homomorphic encryption. With this technique, he says, “you can execute any algorithm, even nonlinear ones, to retrieve an encrypted result.” This opens up a plethora of applications, especially in the realm of confidential Software as a Service (SaaS). Given its potential, it’s no surprise that tech giants like Microsoft are showing keen interest in this encryption method.

However, like all technologies, homomorphic encryption has its challenges. Grieder points out two primary issues: data expansion due to the added noise and the computational resources required, especially during multiplications. Despite these challenges, the potential benefits of this encryption method are undeniable.

Cosmian’s approach to encryption doesn’t stop at homomorphic techniques. They’re also exploring functional encryption. In Grieder’s words, the basic idea is to “encrypt data with a key from which a second can be derived, injected with a function. This key allows this function to operate on the encrypted data and only that function.” The result is obtained in clear text. While there are limitations to the functions that can be executed, many seemingly nonlinear problems can, in fact, be linearized.

Another promising area for Cosmian is Secure Multi-Party Computation (SMPC). This technique allows multiple parties to perform calculations without revealing their data to each other. Grieder provides an example of its application in fraud detection, where companies might want to verify a suspicious case with a competitor. With SMPC, there’s no need for a trusted third party; machines interactively communicate, exchanging encrypted intermediate results. The beauty of this method is that it can be formally proven that these exchanges don’t reveal any actual data used for processing.

Lastly, Cosmian is delving into secure enclaves. While many use them to protect data, Grieder believes their true potential lies in protecting algorithms and bridging them with confidential data. For instance, a startup could send encrypted code to a potential client, allowing them to test it on their data without exposing it.

In conclusion, as data continues to be a pivotal asset in the digital age, the need for advanced encryption techniques becomes paramount. Startups like Cosmian are leading the charge, ensuring that businesses don’t have to choose between data security and innovation. With their groundbreaking approaches, the future of data encryption looks promising and secure.