Cloud Service Recommendation Based On Services Network And Relationships Between Services

Recommendation system is created to solve the problem of information overload on the Internet, recommendation system predicts the user’s individual preferences based on user behavior and recommend merchandise which can meet user’s request to users. Due to the traditional personalized recommendations applies to recommendation of mainly books, videos and other commodity-based, the content which is recommended and the content which user is browsing are often the same kind of products, so the system does not have to consider the type of content or functional properties. In a cloud computing environment, all the resources and information services in any form should be consider when recommend service to the user, the system should know what kind of service is needed, and be clearly what features users want to use, then the system should also consider the user’s interest preferences to complete the final recommendation. This paper focuses on the characteristics of cloud services in cloud computing environment and makes recommendation of cloud services.Combining the characteristics of cloud computing environments users and services and the idea of recommendation algorithm, a personalized recommendation algorithm is proposed based on partnership between services. Partnership between services can help speculate the functional request. When user uses certain services, according to these services’relationships with other services, system will speculate functional requirements and make recommendations according to the relationships. After reach the functional request based on relathionships between services, we will then analyze the user’s personalized preferences. Cloud computing environment using indicators of stability, security, network quality and charges to measure the quality of service (QoS), these indicators just like the describe of the project and key information. System can obtain the user’s degree of concern for different QoS indicators based on the user’s records, and then select service by matching QoS and user preferences.There are a lot of users and services in cloud computing environment, and each user just use only a small part of all the services. Based on the knowledge of complex networks and community, in which, the same or similar activities in the range of users often be seen as a community, in the community users works in close areas and have similar needs for services, rather than make recommendations within the scope of the entire cloud service, the recommendation can be reduced within the range of community and the services involved. This paper presents a community found algorithm in complex network, user community is divided and recommendation algorithm works only within the scope of the community and the services involved. Experiments show that community found algorithm is effectively and recommendation algorithm’s accuracy and speed is improved.