The 6G era is approaching. It will connect the physical, digital and human worlds to provide new experiences, whether in the context of high-resolution mapping for self-driving or remote cars that require microsecond response times to dangers, or the proliferation of mixed reality telepresence. Communications in the 2030s will serve the needs of the future world, embodied by a drive to augment human potential.
While 6G opens a world of digital possibilities, its acceptance will critically depend on a comprehensive 6G security vision, supported by a range of technologies to ensure privacy and trust for the next generation of networks.
The 5G security evolution
5G security design has brought about unprecedented flexibility and transformation. For example, it improved the authentication and key agreement procedure to further reduce the threat of tracking or attacks, where an attacker may be able to verify the presence of a victim device despite the enhancements in subscriber privacy. Additionally, special “lightweight” crypto algorithms that provide high security but, at the same time, minimise the computational effort for low-energy budget devices, will soon be introduced.
Enhancing privacy protection is high on the security evolutionary agenda of not only 6G, but all networks that have come before. 6G requires a developed security paradigm not because it is flawed, but because advancement must be met with protection.
The 6G threat vector
As technology continues to advance, the threat vector broadens. In the context of 6G, the billions of sensors, devices, and new human machine interfaces (HMI) will form the basic threat layer. On the network side, millions of untrusted specialised sub-networks will redefine the arena for attacks, including the risk of malicious appropriation of authentication and identity.
The 6G threat landscape will require five dedicated privacy preserving technologies:
1. Multi-party computation
Multi-party computation will allow multiple parties to collectively perform computation across the cloud, sub-networks and devices, and receive the resulting output without exposing any party’s inputs.
2. Federated learning
Federated learning will allow for flexible and decentralised training of machine learning models at the place where data resides and, for instance, performing training at the edge. Implementing this collaborative approach will first require incentive design to motivate the participation of devices and sub-networks in the federated learning model. Novel federated multi-stage learning protocols will be needed, as well as learning model updates, possibly using blockchain.
3. Data synthesis
Data synthesis is the systematic and controlled generation of artificial data that mimics the dependencies and characteristics of a system’s real data. It is used to extend the data coverage to simplify or transform a model in cases where no real data, or only spare real data, is available. It can be used as a data privacy-preserving technology when, by replacing real data points, it also removes privacy-sensitive dataset features.
4. Homomorphic encryption
Homomorphic encryption allows computation to be directly performed on encrypted data which allows data processing and even generating machine learning models on an untrusted central infrastructure without disclosing the data itself. Once decrypted, the result of the computation matches the result from the computation done on clear text data.
5. Edge profiling
In the 6G world, with large numbers of human-attached sensors, such as ear buds, glasses and cameras, there is a significant increase in the risk of privacy loss because of the inadvertent sharing of private information through these sensors. It is highly desirable to have an automated approach to checking and validating the data integrity from these sensors before data is shared with other applications.
These technologies will ensure that society realises the full value of 6G technology, while creating a resilient, privacy-preserving and trustworthy 6G network.
As we enter a world of increasingly divergent digital standards, governments and policy makers will need to be agile in evolving policies to meet equally divergent industrial and operational security standards needed to support 6G development. Prioritising an expansive global infrastructure of cyber-resilience, privacy and trust as the 6G paradigm evolves will ensure that society can maximise value from this unprecedented networking era.
Michael Montag is the current Head of Cyber Security Research and Dr. Dimitrios Schoinianakis is a Senior Security Researcher at Nokia Bell Labs.
