ASCEND comprises a research line that contains four subprojects: AB-TRAP, SCREEN, CONTROLLER, MALWARE, and DIRECTIONS.
In AB-TRAP, we understand the most common attacks in the reconnaissance phase
and trace them to identify the signature. Thus, it allows us to develop advanced packet filtering IoT firewall. Furthermore, we complement it with anomaly detection traffic, which enables redirect a copy to a honeynet. In this context, AB-TRAP is a comprehensive solution capable of bringing to IoT: adaptive and distributed learning process, unsupervised and supervised malicious identification, model realization in small footprint devices (FreeRTOS, Nuttx, and Linux), and secure OTA.
In SCREEN, we understand how the SOHO Wi-Fi works in the wild at scale by downloading as much firmware as we can. We intend to identify the most common software artifacts (using unsupervised clustering techniques). To reach exploitation to every instance, we conduct a re-hosting task to maximize the coverage amount of original code emulated. Thereby, such a technique advances the state-of-the-art in building realistic and emulated testbeds at scale.
Complementary in CONTROLLER, we offload decentralized algorithms targeted to Intelligent Transportation Systems, typically implemented in the ad-hoc network. Here, such Network Functions are Virtualized in Multi-access Edge Computing nodes for the next Mobile Network Generations (5G, 5GB, and 6G). Our approach consists of building a 5G-core and implement URLLC on top of it. In addition to that, we can exploit realist threats and build more robust protection mechanisms. Containers and virtualization correspond to the main research interests.
MALWARE subproject concerns identifying ransomware behavior since the initial access, passing throughout the persisting in the network hosts to the effectively dispatching the attack. We collect the most common ransomware families, trace their execution on logs, and analyze them statically and dynamically. Understanding the whole chain of events, we can provide protection and prevision of ongoing threats. Anomaly detection plays a crucial role in this context by helping to identify zero-day behavior.
Understand the propagation of malware in the darknet is also useful. To this end, the DIRECTIONS subproject aims to understand the dynamical behavior of malware propagation, which includes coordinated actions—mapping the relationship among the actors in the propagation of malware. Our approach employs complex networks to extract interaction backbones and plot evolutional events chains leading to infection. Therefore, we can identify, characterize and develop protection mechanisms against new threats.
... [more]