Managing Network System by Bloom Filter

Abstract

Network management system (NMS) is important both in ensuring the correct operation of network devices and in maintaining the services that run on them. However, the relentless growth of DSL's users, meaning that IP traffic volumes nearly double every two years, renders real-time monitoring and analysis of every customer's service parameter is a very challenging problem. The most important problem is the bottleneck in centralized NMS which most of services provider use as SNMP-based system. This paper presents a new model which will be able to ease the problem from bandwidth consumption of SNMP-based by two main methods. Firstly, to ease the problem from bandwidth consumption of SNMP-based, this model will use the probabilistic data structure to decrease the number of packets of monitoring network device's parameter. Secondly, to ease the problem from the relentless growth of the number of the network's devices, this model is designed to support distributed network devices' operations in a distributed NMS fashion. Meaning that, the number of NMS's devices can be increased with the number of network's devices if it is necessary. By using both of two methods, customer expectations with bandwidth requirements and reliability requirements can be improved with this model.