Proactive Recommendation of Composite Services in Multi-Access Edge Computing

Zhizhong Liu, Quan Z. Sheng, Dianhui Chu, Xiaofei Xu, Hedan Zheng, Kai Feng

Abstract: Multi-Access Edge Computing (MEC) is an emerging computing paradigm that brings services from centralized cloud to nearby network edge to improve users’ Quality of Experience (QoE). With massive services from different domains being emerging in MEC, various powerful composite services can be created with simple services to satisfy users’ complex needs. However, existing service composition methods follow a passive service model and cannot proactively recommend optimal composite services to users in MEC, which seriously affect users’ service experience. To tackle this issue, we propose an approach for proactive recommendation of composite services based on demand prediction. Our approach consists of three steps. First, we predict a user's service demand based on an attention enhanced deep interaction network (AEDIN) model trained with clustered data. Then, we create the optimal composite service to satisfy the predicted demand with a mobility-aware services composition method, and finally, we proactively recommend the optimal composite service to the user. The extensive experiments have been carried out to verify our proposed approach and prove its performance superiority.