Font Size: a A A

Bottom-up Service Identification In SaaS Migration

Posted on:2020-07-17Degree:MasterType:Thesis
Country:ChinaCandidate:T WuFull Text:PDF
GTID:2428330596992468Subject:Software engineering
Abstract/Summary:PDF Full Text Request
The growth of software scale makes software development more and more difficult.SaaS migration is considered an important means of software development in the cloud environment due to its flexibility and scalability.In the process of migration,the part of legacy system or software needs to be packaged as a service for software development,so how to identify a complete and independent service is very important in the migration process.Based on the problems of low service quality and poor recognition quality in existing service identification methods.The thesis improves the software clustering method based on class relationship type information and optimizes the process of bottom-up service identification.The thesis mainly has the following innovations:1.The software clustering method based on class relationship type information is optimized.A semi-random initial population generation method is proposed.An adaptive genetic algorithm is introduced to generate crossover and mutation probability according to the characteristics of the population to ensure the diversity of genetic algorithm.2.The multi-objective software clustering problem is transformed into a singleobjective problem by combining subjective and objective weights.The experimental results show that the optimized single-objective clustering algorithm can not only ensure the objectivity of the algorithm,but also shorten the clustering time significantly.3.The workflow-based cloud resource selection system is taken as the test.The improved algorithm is verified by the system.The cloud resource selection system is improved according to the analysis results,so that the system has better portability.The thesis verifies that the improved method has improved obviously in the accuracy of service identification and time efficiency through a large number of comparative experiments,which guarantees the quality of SaaS transplantation.
Keywords/Search Tags:SaaS migration, service identification, software clustering algorithm, multi-objective optimization
PDF Full Text Request
Related items