Application of Category Theory to Network Service Fault Detection

Network automation has become crucial in supporting services in 6G networks. This mainly derives from the complexity of the composition of numerous distributed virtual network functions (VNFs) in creating highly flexible virtual network services. Therefore, a network service automation system is a key technology enabler for 6G. However, the added complexity renders network service automation systems particularly sensitive to faults, some of which cause network outages that harm the smooth operation of basic societal services. Current state-of-the-art (SotA) solutions for fault detection can barely detect hidden faults. Herein, we propose a mechanism for automated network service analysis (ANSA), which constructs and analyzes a digital twin of a network service. The digital twin represents the available information about the network service based on category theory. It uses the properties of category theory to perform an analysis through which the faults of the network service are identified. We evaluate a prototype of a network service automation system that incorporates ANSA to demonstrate 1) the benefits of using digital twins for analyzing network services, 2) the benefits of using category theory for constructing digital twins of the network services, and 3) the resulting improvements in fault detection. Overall, ANSA can detect an average of 94% of the faults present in a network service. In comparison, previous SotA solutions can detect only 30%–50% of all faults.