Scalable Solution for Service Function Chaining Embedding in Hybrid Optical-Electronic Networks

Network function virtualization (NFV) leverages software and general-purpose computation devices to provide various network services through service function chaining (SFC), where data flows through a sequence of network functions to obtain the desired services. In this context, the growth of network traffic due to the proliferation of network services presents a challenge. To address this, hybrid optical-electronic networks, combining the advantages of both optical and electronic networks, have been applied to create a more efficient and flexible communication system which also transport traffic through the network functions of SFC. In this paper, we address the problem of embedding SFC in a hybrid optical-electronic network. This problem is even more challenging than the SFC embedding problem in an electronic network, which is already known to be NP-hard. To tackle this problem, we propose a scalable SFC embedding algorithm in hybrid optical-electronic networks that optimizes both optical and electronic bandwidth allocations as well as computation resource allocation. We first provide an integer linear programming formulation for this optimization problem and then propose a scalable heuristic algorithm based on the randomized rounding method to achieve near-optimal solutions. Our numerical results demonstrate that the proposed algorithm achieves both guaranteed performance and scalability through efficient running time.