Performance Analysis and Optimization for Mobile Agents in Evolutionary Network
Zhongzhi Luan · Xi'an Jiaotong Daxue xuebao · 2010
A novel fast launch algorithm is proposed to improve the performance of mobile agents in evolutionary networks.Analyses on the performance data of mobile agent's non-execution forwarding,execution forwarding,and detail execution steps show that the execution environment preparing process for mobile agents is the bottleneck.When the fast algorithm is executed,all codes in the mobile agent package will be extracted and loaded prior to the mobile agent's main program.Before a new code is loaded,the preloaded codes will be searched,and hence the code image can be located quickly.The mobile agent's execution forwarding can be improved by optimizing the searching order.Testing results show that the average loading time of the mobile agents is reduced by 3.26 ms and the throughput of network nodes is increased by 58%.