A service cache-based dynamic collaborative task migration technology

The continuous emergence of compute-intensive and delay-sensitive services in edge networks has promoted the rapid development of task migration technology. However, task migration faces significant technical challenges, such as complex application scenarios and the high complexity of problem modeling, particularly when accounting for user mobility. Therefore, a task migration framework integrating optimal clustering strategy and edge pre-caching is proposed to ensure the stability and continuity of user services in mobile-aware environments. Firstly, the concept of dynamic cooperation cluster and migration prediction radius is introduced based on user movement trajectory, and then, a pre-migration model is proposed to determine when and where to migrate, with a specific focus on two task scenarios: mobile and overload. Under the constraint of maximum tolerable delay, the limit value of cooperative cluster radius and target server quantity are derived. Finally, a user-centric distributed multi-server cooperative clustering algorithm and a service cache-based Double Deep Q-learning Network are proposed to effectively address the issue of optimal clustering and service caching. The experimental results demonstrate that the proposed migration selection algorithm can achieve a minimum reduction of 16.46% in migration cost and a 12.55% decrease in delay compared to other algorithms.