Research Of Initiative Management Scheme For Sensor Networks Based On Data Aggregation

Along with the development and maturation of wireless networks andcommunications, the research of WSN on industrial application and academicstudy area is getting a great development.However,sometimes because of theimpact of un-expectable environmental factors or errors on interior system, thedeployed WSN may run with much significant deviation. And therefore, it isnecessary and important to improve the network management skills.According tothe characteristics of WSN,management system should make the practicalapplications can be in motion more reliably and effectively, and with robustrecovery mechanism to against most malfunctions. The network managementinvolves network performance monitoring and network recovering. Networkmonitoring plays an important role in managing system, it helps discovering theinvalid nodes, and records the energy and communication expense of network andother error information. How to effectively and accurately capture the data ofevery hierarchy is a base and essential study point.For the aims that improving the energy controlling and initiative managingproblems, in this paper we present a initiative management scheme, which uses anamed tactic middleware model combining with Mobile Agent to get amelioration.This scheme constructed and designed a middleware plat form, whichbased on tactic agent module, to make a abstraction and calculation on datamonitoring and aggregating process, combining with mobile agent taking ondistributed computing. It establishes a delaminated initiative managementmechanism through aggregation tactics, which departs the data aggregatingprocess into3asynchronous disposing phases that consist of Pretreat, Scan andDiagnosis.basing on the data aggregation,through setting up a tactic managinginformation bank to institute explicit supervisory control strategies for networkmonitoring,it establish a named relating monitoring tactic,that can build a relationin which the variety of network state is associated with the variety ofnodes’parameters. This scheme detaches the performance supervising of basicrunning parameters from the monitoring of action and business of network,andestablishes a group of quantized and standardized uniform mappingrelations,aiming at enhancing the pertinency and timing-effect on control andmanagement in administration of networks.The final, paper has tested and verified the availability and effect of thismodel on energy controlling, cluster-header maintain period length and theresponse time of task dealing, through a simulation experiment running onOMNet++.