Abstract: Under virtual computing environment(VCE), tasks have features of large quantity, ambiguous requirements, and various types. This makes resource matching face enormous challenges. According to the data published by VCE platform, a resource matching optimization model combined with resource and task clustering was proposed. By analyzing task requirements and consumption characteristics, the model clustered tasks based improved bised K-means, and combined with the virtual resource types to generate optimized matching resource list. The experimental analysis verified that the model effectively reduced resource matching range and improved the successful rate of tasks to provide the foundation for precision matching.

Key words: task clustering, bisect K-means, resource matching, virtual computing environment