An Analytical Model for Centralized Service Discovery Architecture in Wireless Networks

Service discovery (SD) has been a hot research topic for several years, while few efforts have been made in theoretical analyses. Only one paper concerned the modeling of the decentralized SD in Ad-hoc networks. However, the traffic load caused by broadcasting restricts the adoption of decentralized architectures across the Internet. The purpose of this paper is to propose an analytical model for the centralized SD architecture in wireless networks. The directory is modeled as an M/G/c/c queuing system, which evaluates the cache usage on the directory. Moreover, we study the periodic service announcements (SAs) of an individual service provider (SP) to derive the success rate and the produced network traffic. Particularly, we take into account the characteristics of radio links and nodes' movement. The numerical analyses show that the update interval influences the metrics incompatibly. Trade-off between the success rate and the overhead is the only way to optimize the system performance. We also observe that nodes' movement can only cause degradation in success rate. And, the degradation enlarges with the increasing update interval.