Font Size: a A A

Researches On Key Technologies In Cloud Computing Resource Management

Posted on:2016-10-20Degree:DoctorType:Dissertation
Country:ChinaCandidate:M LiFull Text:PDF
GTID:1108330473452479Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Cloud Computing is the development of distributed computing, parallel computing and Grid Computing. Meanwhile it is the commercial implementation of above scientific concepts. How to manage effectively the dynamic, heterogeneous, distributed, and autonomous Cloud Computing resources becomes an important Cloud research domain.Based on the research of current relevant works on Cloud Computing resource management, this dissertation focuses on four key technologies in Cloud Computing resource management such as Cloud resource management model, Cloud Computing resources organizational model, Cloud Computing resource discovery strategy and Cloud Computing resource schedule strategy. The major contributions of this dissertation are as follows:(1) This dissertation proposes a multi-agent system-based Cloud Computing resources semantic organizational model. It utilizes the intelligence, distribution, asynchronous, and autonomy of multi-agent system to manage the dynamic, heterogeneous, distributed and autonomous Cloud Computing resource. Based on the multi-agent system-based Cloud resource management model, this dissertation inducts clustering to classify the Cloud Computing resources based on their semantics and organizes them in a tree manner. It is easy to add new Cloud resources and it has low bandwidth in such organization manner. The simulation results show the availability of this model. And it can solve the problems of poor scalability and overload bandwidth etc. in existing Cloud Computing resources organizational model. Hence it provides a good organization manner to the Cloud Computing resource organization.(2) This dissertation proposes a semantic search engine(SSE)-based Cloud Computing resource discovery strategy. Based on the multi-agent system-based Cloud resource management model, the strategy inducts SSE to complete the Cloud Computing resource search. The core of the approach is ontology-based resource discovery strategy and the strategy uses ontology to descript "Request-Resource" asymmetricly and semantically, hence it can effectively optimize matching between Request and Resources. Futher the construct of the ontology is based on a set of rules. The simulation results show the availability of the Cloud Computing resource discovery strategy. This strategy can effectively solve the low efficiency and poor reliability in current Cloud Computing resource search methods.(3) This dissertation proposes an improved parallel genetic annealing algorithm(IPG2A)-based Cloud Computing resource schedule strategy in the multi-agent system-based Cloud Computing resource management model. The strategy inducts the basic theory of genetic annealing algorithm and fully integrates the quick global search capability in GA and local search capability in simulated annealing algorithm to make IPG2 A a wide application in Cloud Computing. And IPG2 A can increase Cloud resources scheduling efficiency. Simulation results demonstrate the availability of the IPG2 A and show the Cloud Computing resource schedule strategy based on IPG2 A can solve the single point of failure, the communication overhead and low scheduling efficiency in existing resource scheduling.
Keywords/Search Tags:Cloud Computing resource management, multi-agent system, clustering, semantic search engine, genetic annealing algorithm
PDF Full Text Request
Related items