Design And Implement Of Distributed Probe Scheduling System

With the increasing scale of network,the scheduling method of probe nodes must be improved to measure the performance of large-scale network.However,if we select all the probes to perform network measurement,it will cause large-scale network measurement traffic and bring high consumption to the network measurement platform and probe nodes.In addition,the probe status will have a huge impact on the network measurement results.When the probe load exceeds a certain limit,downtime or other failures may happen,resulting in the reduction of measurement accuracy.In the existing network performance measurements,the scheduling of probes mainly focuses on the coverage of the region or autonomous system,but pays little attention to the status information of probes and the number of probes selected.To address the above problems,this paper proposes a multi-objective optimization model that takes the coverage of the target network as a constraint,and optimizes the number of probes and the probe load under the condition of satisfying the desired coverage.Based on the above multi-objective optimization model,this paper designs a probe scheduling system,which performs network measurements on the target network by selecting some probes,and achieves the reduction of probe resource consumption and overall probe load balancing while being able to obtain real-time service status of the target network.The system mainly includes modules such as task creation,probe scheduling and measurement result query.Among them,the task creation module is responsible for providing users with a visual task creation interface when they create tasks.The probe scheduling module is responsible for generating probe sets for users.At the same time,the probe scheduling module also can monitor the probe status and replace the faulty probe in time when the probe fails to avoid the measurement results being affected.The measurement result query module is responsible for displaying network measurement results to users,from which users can obtain detailed information about alarm events,so that faults in the network can be fixed in a timely manner.According to the results of the final system test,under the circumstance of meeting the conditions for coverage of the target network,the probe set selected for the probe scheduling system designed in this paper is at least 5%lower in the number of probes than the original system.Furthermore,it is at least 14%lower than the original system in terms of probe load.