Fast Failure Detectors for 5G Architectures
Advances in the fields of Machine Learning, AR and VR have opened up avenues for a lot of exciting applications ranging from IoT to real-time active guidance. A large majority of these novel applications fall under the category of Edge Native Applications, and hence cannot fulfil their functionality without an Edge-based network architecture as a result of their tight latency budgets requirements. Edge-Computing is dictating the design considerations for next-generation cellular networks such as 5G, and their control planes to cater to these low-latency requirements. While research efforts have produced results in the form of modified cellular control plane designs that focus on reducing delays and masking failures in the control plane, the fast detection of such failures remains a challenge. Existing failure detection techniques employ centralized approaches such as keep alive messages or echo messages from one component to another and are prone to delays due their dependence on a single network link for communication and a single observer for failure detection. In this work, we present a Distributed Heartbeat algorithm, where Control Plane Functions (CPFs) exchange heartbeats in a round-robin fashion and participate in the failure detection process. We evaluate the proposed scheme through simulations and present the promising results.
Join Zoom Meeting Link: https://zoom.us/j/94246773753?pwd=UXdPd0NUUzhBMGNvWDIvZ1RNME5pZz09
Meeting ID: 942 4677 3753