Security, Privacy, Safety, and Ethics (Trustworthiness) are essential today in telecommunications due to the proliferation of massive DDoS attacks and there is a need for definitions and regulations related to how to deal securely with user data. SMARTNESS approach for trustworthiness is considering the protection of networks against intrusions, the guarantee that the AI decisions will not deteriorate a network, and that all the proposed mechanisms will respect rules introduced by recent data protection regulations. Besides, the possibility to control cyber-physical systems (vehicles, robots, etc.) acting in the real world with low latency in 5G and post-5G network architectures will be considered, since it increases the criticality of network vulnerabilities. In this scope, the following topics will be investigated to guarantee a secure and trustable network ecosystem:
- Security, Privacy, and Trusted-by-design: Each new service integrated into a large scale system, such as a communications network, has the potential to increase the attack surface of such a system. Keeping a small attack surface requires a regular routine of system scans, considering both classical vulnerabilities and new ones usually disclosed in public services like the National Vulnerability Database [NVD, 2020]. To avoid the interruption of production services, it is necessary to design them considering that vulnerabilities can arise and, if this happens, updates will need to be applied transparently;
- Early Warning Systems based on Machine Learning: The appearance of security threats in a communications network, both by misconfigurations or by discovered vulnerabilities, can be early detected if there is a pattern in the network traffic considered as normal. To infer, and update, this pattern requires the deployment of Early Warning Systems with Machine Learning capacities. These systems could raise alarms and, in automated and programmable environments such as SDNs, install flow rules restricting suspicious flows in the network [Batista et al., 2016];
- Secure identities: Future services in mobile networks can involve several partners to work (cloud providers, edge providers, highway concessionaires, etc.) and it is important to keep track of the user identity in several interactions between these partners to ensure the correct permissions and the contracted QoS. In this scenario, the authentication of the user identity is an open issue. Solutions brought by SMARTNESS will require innovative architectures that could, for instance, employ Blockchain [Gorla et al., 2020];
- Trustworthy AI: Artificial Intelligence (AI) mechanisms are crucial to the next generation of the Internet and these mechanisms need to be trusted by both providers and users [Spada and Vincentini, 2019]. Low rates of false positives and negatives during the classification of DDoS attacks and protection against data with biases during the training phase of machine learning algorithms used to predict the network usage are some examples of required features to these mechanisms. SMARTNESS members will keep up to date in recent AI topics to integrate the most promising in the proposed solutions;
- Efficiency of cybersecurity mechanisms: The future Internet is expected to be the Internet of Everything. In this network, several small and low-powered devices will take advantage of the processing capacity of edge and cloud to make the best decisions to solve several problems. Although it is possible to deploy cybersecurity services remotely, some basic cybersecurity mechanisms (e.g. cryptographic protocols) still will need to run locally in the devices. These services need to be optimized to avoid a significative processing delay in communication and excessive energy consumption [Binti and Hassan, 2019] specially when they have to adopt new quantum-resistant cryptographic algorithms;
- Regulatory compliance: The rules introduced by the General Data Protection Regulation (GDPR) and by the Brazilian General Data Protection Law (LGPD) [Koch, 2019] need to be considered when data, from different users, are being moved to be processed in edge and cloud infrastructures. All the solutions pursued at SMARTNESS will take this in consideration in order to guarantee the privacy of the users when storing data and when inspecting header and payload of the packets;