Application Store Based On Cloud Computing Community Systems Research And Development

Last 3 years,social business software in enterprise was in the momentum of rapid development and this trend will continue. In the end of 2011, there will be more than 40% of SMEs who advertise for using the social network. SNS (Social Networking Service) will infiltrate and become an important part of enterprise applications. Enterprises must have a platform of their own SNS. However, for SMEs, the introduction of SNS system faces three difficulties.1) One-time investment costs and maintenance costs are high.2) User scale is not fixed and the number of uesers is fast growth.3) Time is concentrated, but not fixed for users access.In order to solve the challenges in front of the three questions, author proposed SNS cloud computing in this paper. The cloud computing model is designed for the characteristic of SNS system. on the basis of Eucalyptus, It add a batch message processing server clusters and server clusters to handle the SNS system batch tasks and publish / subscribe type of function.The purpose is to reduce the load on application servers and to improve response speed. Also in front of the three difficulties for the last one, dynamic resource scheduling automatically problems.The author proposed a solution and the main algorithm. Solution to this problem is divided into three steps.1) Hot probe——the real-time detects each running virtual machine's memory, CPU utilization and network bandwidth and other performance parameters are exceeding the prescribed threshold, if more than threshold and the resulting is hot spots.2) In the data center it finds qualified targets resources, first look for the server cluster which produced in the hot. if no, then to the other server clusters to find,.until you find the eligible target resources.3) At the target resources launches a new virtual machine, deploy systems of hot spot, shared access request with the hot spots. the original hot spot virtual machine's memory, CPU usage and network bandwidth and other performance parameters will be restored to the normal range, that is, to eliminate hot spots.The SNS cloud computing model is designed to sovle the three challenges. which fully take into account the SNS system features, make the SNS system on the cloud computing platform has good performance and resource utilization is high.And one of the core issues of dynamic resource scheduling automatically proposed solutions and the main algorithm.The second half part on the basis of"Changhong application store SNS community system", based on the use of SNS cloud computing model, design and implement application software store community system. to verify the feasibility of the SNS cloud computing model.