Font Size: a A A

Research On Technology Of Web Services Fault Management In SOA

Posted on:2014-02-05Degree:MasterType:Thesis
Country:ChinaCandidate:P S CuiFull Text:PDF
GTID:2308330479479154Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
As a solution to make the software loosely coupled,Service-oriented Architecture(SOA) has attracted wide attention from industry and academics. Especially after the maturity of XML and SOAP technology, the Web Service appears and SOA is widely used. In this context, the paper focuses on service fault management, especially management of composite service. Service fault management can timely receive the state of service, monitor the symptom of service and locate the fault to ensure that the critical service can run well. In this condition we can improve the service’s QoS. SOA brings new features, and the current fault management technology has drawbacks and cannot apply SOA very well. To solve the problem, the paper makes study in composite service fault management, and has mainly completed the following aspects:(1)The model of composite service fault management. The model defines the managed objects, the content and function of fault management. We can locate the fault in the resource which the service depends on by collaborating with other management systems. For the challenge of mutil-domain case and virtualization technology, we design the fault management model for mutil-domain and virtualization environment.(2)The fault location technique. A hierarchical model is proposed, which can get symptom by using active probe. We also design the probe selection algorithm based on the hierarchical model. There exists noise in the internet, so the service’s state may be uncertain. In this situation, we design a fault location algorithm based on Bayesian network. We also propose a fault location algorithm using the indirect dependence. The simulation result shows the effectiveness and efficiency.(3)The analysis and forecasting of fault. After the fault is located, we need to assess the impact of fault and predict which service will come into the fault state. The paper presents an algorithm to calculate the impact of faults, and also an algorithm to predict which service may come into fault state.(4)Finally, we design and implement a prototype of the service fault management system to verify the technology. The service fault management can effectively monitor the service’s state, locate the fault accurately, and intuitively analyze the fault’s impact.
Keywords/Search Tags:SOA, combinded service, fault location, analysis of fault impact, fault prediction
PDF Full Text Request
Related items