Design And Implementation Of Open Source Active Detection System Based On Automation Template

The worldwide and continuous network measurement is an important way for the design of high performance protocol,network planning construction and problem locations.Due to the interface openness,wide distribution and the rich measurement tools,Open source(Looking glass)becomes the infrastructure to support network measurement at the long-time granularity across domains and layers.However,the breadth and depth of its usage is limited due to the various interfaces,the variable probing tools and the low-frequency probing.It is significant to improve the usability and extendibility of open source.Firstly,this article performs the systematic analysis and classification for the open interfaces and introduces an active probing system based on automatic template,including the autonomous discovery of open source,the auto generation of task scheduling module,and the flexible assignment,which implements the automatic and uniform invocation.Secondly,this article develops a new algorithm for evaluating the efficiency of the open source.The new method performs comprehensive quantitative assessment for the availability of open source,which is based on the periodic active probing and the Bayesian inference model.A lot of experiments have been performed and it turns out this system is of significance to construct the low-cost and sustainable distributed network measurement infrastructure.