In a bid to efficiently realize that joint-protecting privacy has a key to the fastest computing solution, the Chinese multinational technology company Alibaba is set to develop a product based on privacy computing.
Recently, the Cheetah secure two-party computing framework newly developed by the Chinese tech giant Security has greatly improved the overall performance of two-party computing.
The quickest computing solution is more than 5 times quicker than the world’s greatest computing solution, Microsoft CryptFlow2. It’s used in risk management, and the relevant research findings were accepted by the USENIX Security Symposium 2022, one of the top four worldwide security conferences.
The “millionaire conundrum” is a common two-party calculation scenario in which two millionaires want to compare who has greater wealth but neither wants to reveal their wealth estimates to the other. Complex and expensive cryptographic algorithms must be constructed in a targeted manner to achieve this security aim, resulting in massive ciphertext volumes and poor processing speeds.
Hong Cheng, a senior security expert at Alibaba Security, explained that, in the case of an image recognition service that protects user privacy, if customer A has a photo, server B must perform AI recognition on the photo to determine whether it contains non-compliant content, but cannot view the photo due to privacy concerns.
It takes hundreds of seconds to complete an image identification using Microsoft’s CryptFlow2, which was the best before, whereas “Cheetah” makes the speed 5 times faster and can be done in very few seconds under the premise of ensuring the same verifiable security., a major step forward from practicality.
The industry and the financial community have been paying close attention to privacy computing technology. Last year, related startups raised approximately 200 million yuan in Series A funding, breaking the previous record for a single-round A round in the privacy computing space.
He believes that an open-source “Cheetah” that can achieve true provable security will assist the industry in better understanding the current state of privacy computing, defining best practice guidelines for safe two-party computing, and encouraging the healthy development of privacy computing technologies. The first-generation security architecture has arrived in a field that creates security systems from the ground up.
In his explanation of about the computing measures for standards security best practices, Hong Cheng said:
“At present, my country’s privacy computing industry lacks the measurement standards and best practices for ‘security’, and some users are more exposed to non-provably secure privacy computing solutions represented by federated learning.”
“There is a lack of awareness of the importance and difficulty of security, so it is often wrong to overestimate the ability of privacy computing and underestimate its difficulty. For example, we have previously found some security vulnerabilities in the encryption algorithms used in open source federated learning frameworks.” He added.