The Overload Control Algorithm For Multi-SCP Intelligent Network With Priority

The capacity of the SCP(Serviece Control Point) is considered to be the brief bottle neck of the Intelligent Netowrk. Establishing an efficent Intelligent Network SCP overload control mechanism can contribute to the stability of Intelligent Networks so that it could provide high quality serviece. Traditional Intelligent Network overload control algorithm is mostly oriented to single SCP Intelligent Network environments. Accompanying the developement of the Intelligent Network servieces, Intelligent Network's architecture has been developed from single SCP to multi-SCP, and formed two modes. The research of overload control in AIN(Advanced Intelligent Network) has been unfoldered earlier. Because of the complexity of multi-SCP Intelligent Networks, its overload control research is very little. The traditional reasearch results of SCP overload control in single SCP Intelligent Networks or AIN both cannot be used in multi-SCP Intelligent Networks directly. Therefore, designing a logical and efficent SCP overload control algrithm seems very necessary.The architecture of distributed multi-SCP Intelligent Network requires that the Service Switch Points (SSPs) must have the capability of service application routing. This paper introduces the vector load index in a computing cluster to the multi-SCP Intelligent Network system, and proposes a mechanism which is based on vector load index to solve the SSP'srouting problem in this environment. At the same time, the vector load index is used as the key parameter overload detecting algorithm too.With the development of Intelligent Network serviece, absolutly fairly serviece call application method is impractical. The concept of"Priority"is introduced. The paper presents how to deal with the serviece call applications with priority in a distributed multi-SCP Intelligent Network.This paper proposes an SCP overload control algorithm based on ACG(Automatic Code Gapping) for distributed multi-SCP Intelligent Network enveronment. The simulation experiment results show that the mechanism proposed in this paperis is simple to be realized and efficent.