Trust-Aware Service Chaining in Mobile Edge Clouds with VNF Sharing

Network Function Virtualization (NFV), characterized by implementing hardware network functions as software in virtual machines (VMs), is a promising technique to provide flexible and agile network services. Meanwhile, mobile edge cloud (MEC) is envisioned as an enabling technology for providing extreme low-latency network services by deploying cloudlets within the proximity of mobile users. However, it is challenging to find an optimal placement for VNFs when considering service chaining requirements in an MEC with cloudlets having limited computing capacity and users having stringent delay requirements. In addition, security requirements of network services in the MEC makes the problem even harder. In this paper, we investigate the problem of trust-aware throughput maximization problem in an MEC, by incorporating QoS, trust requirements of user requests and allowing VNF being shared among mobile users. We first devise an exact solution to the problem by formulating an Integer Linear Program. Due to the inscalability of ILPs for large problem sizes, we propose an efficient heuristic algorithm based on a randomized rounding technique. We finally evaluate the performance of our proposed algorithms against the state-of-the-art by simulations. Results show that the proposed algorithms achieve higher system throughput than their counterparts by 10%.