Queue-aware reliable embedding algorithm for 5G network slicing

Network slicing has been proposed as a promising technology to fulfill users' specific service demands in 5G network. The important issues in this network paradigm are slices' reliability and online network slice embedding (NSE). In this paper, an NSE model under the reliability constraints is established based on cloud radio access network architecture. The model jointly maximizes the number of network slice requests and minimizes the failure rate of NS. We further investigate the tradeoff between network stability and NS reliability. The NSE problem is formulated as a mixed integer linear program (MILP) and the heuristic algorithm named the queue-aware reliable embedding algorithm is proposed to improve the reliability of NSE. The Lyapunov optimization model is introduced to optimize the resource allocation while ensuring the stability of the queue. Furthermore, an NSE-online mechanism based on time window is proposed that can realize the online processing of the incoming NSRs. Both theoretical and simulation analyses are conducted and the results demonstrate that the proposed algorithm can improve the throughput while guaranteeing the network reliability and stability.