Key Technology Research On Cloud Storage Service Selection

With the rise and development of cloud computing technology, companies are gradually changing their acquisition and use of IT resources and services. More and more companies tend to migrate all or part of their business processes to the public cloud computing environment, or to wrap up its excess IT resources into public cloud computing services leasing to other companies, thereby reducing their operating costs. The rapid spread an d popularity of cloud computing, resulting in a lar ge number of cl oud computing providers, also known a s cloud service providers, offer a variety of different public cloud services, these cloud services often have different abstract level, quality of se rvice and pricing policies. Thus, on one hand companies have more choices, but on the other hand, they can’t make a good choice. Namely it’s a question how to find the best fit of their business processes and cost-effective service from more and more cloud services with different characteristics.This paper based on the urgent needs of the enterprise, focusing on cloud storage services, research the key technology of cloud storage servic es selection. F irstly, cloud computing, infrastructure as a service, cloud storage service and service selection techniques are outlined, then the non-functional attributes also known as Quality of Service, w hich determine cloud storage service selection are de tailed analyzed, the features and acquisition m ethod of QoS attribute data are listed, QoS indicators system are constructed. A QoS-based cloud storage service selection framework and the component module of the fram ework are designed and detailed analyzed. Recommended Rules Module generates recommend rules based on user needs; Users Rule Module receives and feedbacks user rules. QoS Testing and Collec tion Module collects and standardizes QoS property values of the cloud storage service providers to provide the date for the Match and Filter Modul e and the Integrated Sorting Module; Match and Filter Module ranks the cloud storage service providers according to the long-term history QoS data sorting by different recommended rules and selects the top 50 service providers into the candidate list, thereby reduc ing the computational complexity of the Integrated Sorting Module. Integrated Sorting Module does a comprehensive sorting of the cloud storage service providers in the cand idate list in a ccordance with the user rules and short-term QoS data. Feedback and Correction Module modifies the recommend rules according to the selection and feedback of users and the feedback correction algorithm recommendation rules to m ake the recommend rules meeting the user needs better.Finally, the paper carried a prototype im plementation of the fra mework and makes the appropriate simulation to perfect the recomm end rules base. The paper does a series of experience based on the typical cloud storage services in the market, and then makes a compare and analysis of the services’ performance to provide a reference for user’s selection.