Font Size: a A A

Feedback Enabled Runtime Behavior Monitoring Methods For Distributed Software

Posted on:2013-02-12Degree:MasterType:Thesis
Country:ChinaCandidate:J HuangFull Text:PDF
GTID:2248330395462137Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years, with the improvement of software developingtechniques, the distributed software plays an important role in the medicaltreatment, traffic, military affairs and many other important areas. Due tothe characteristic of the distributed software, such as huge scale, highcomplexity and openness, though it is widely adopted, the problem of itsfailure and fault gradually increased. People are eager to monitor thedistributed software more effectively and make it operate more safely andreliably.In order to realize the goal of adopting the feedback results to adjust theinteractive behavior monitoring, in the paper, the architecture of themonitoring system is proposed, and the design of each functional module,the realization of the monitoring system and the performance testing arealso discussed.The monitoring system design criterion based on the feedback-enabledinteraction behavior is formed. According to the criterion, a multi-laymonitoring architecture is designed, which including monitoringinformation dissemination, monitoring analyzer, monitoring intentionmanagement, component information display, monitor componentsdeployment, and various functional modules are investigated.To divide the interaction of functional modules, the paper firstly clearsthe monitoring object, which helps to accurately monitor interactivebehavior and record corresponding monitoring data. Monitoring intentionmanagement method research is to make more accurately expressmonitoring requirements and more efficient control scale and strengthwhich belong to distributed software interaction monitoring. The display ofcomponent information makes use of the reflection technology to acquireinformation that internally encapsulated in components. The display makesuser more intuitively understand monitoring objects. Using complexnetworks to describe the software behavior and quickly build a model withcomplex network, monitoring data and behavior matching, so that thecurrent behavior of the runtime software can be flexibly judged, analyzed,matching rules already enacted, the current scale of monitoring, thestrength to make adjustments. To analyze the results of feedback to enable the monitoring of behavior, in order to adapt the change of environmentand adjust monitoring demand in predefined rules, the system have to adaptto monitoring environment change. The monitor deployment device wasdesigned to configure the monitoring demand and dynamic adjustment ofdemand, so as to achieve feedback-enabled monitoring. Transmissionnetwork limited bandwidth, so it can’t satisfy huge amount monitoringinformation transmission requirements, the serious influence to promote theefficiency of the monitor. In this paper, we establish the systemperformance data update storage manner. Fully utilize the transmissionnetwork, and improve the quality of monitoring information acquiring,easing performance bottleneck.The major components of the feedback-enabled monitoring systemwere realized, performance testing was also carried on, and verified thesuperior compared to other monitoring methods.
Keywords/Search Tags:distributed software, monitoring method, feedback-enabledmonitoring, self-adaptive, rule, adjustment
PDF Full Text Request
Related items