A dynamic delay-based reliability evaluation model for communication networks

Abstract

The traditional network reliability research generally focuses on the topological connectivity among given nodes based on graph theory. Without considering the relationship between network failures and network performance degradation, traditional network reliability analysis cannot reflect actual network ability reaching up to a certain accomplishment level over a utilization interval. As for the performability analysis of network with specific requirement of transmission time, this paper proposes a dynamic delay-based network reliability evaluation method, which considers network’s ability that transmission delay can be controlled under expected upper limit level in the course of network normal operating period, and puts forward a three-layer network reliability analysis architecture. By integrating a network component reliability model, a network topology model and a network performance model, the dynamic delay-based network reliability model is established, and the network reliability influence factors are integrated into network reliability evaluation. A simple example shows that for a given network, network reliability is different from specified performance requirement. Furthermore, network failures will result in network performance degradation in network long-term operation. It is also proven that network reliability relates not only to network topological connectivity but also to network application environment and specified performance requirement. The key contribution of this paper is to guide network engineers to design a high performance network by optimizing the network parameters according to the delay-based network reliability evaluation method proposed in this paper.

Keywords:

• Delay-based network reliability evaluation
• Three-layer network reliability analysis architecture
• Queueing model
• Communication networks

Additional information

Funding

We would like to appreciate the supports of the National High Tech Program 973 Project (2014CB046402) and National Natural Science Foundation (51175014).