Virtualized Resource Image Storage System Based on Data Deduplication Techniques

With the rapid development of virtualization technology, OpenStack and Kubernetes as virtualization service have gradually become the core content of cloud computing. Virtualization technology is transforming to the direction of containerization. Many enterprise users are gradually increasing their support for container images while using virtual machine images. The number and scale of virtualized resource images have increased significantly. For example, there are currently more than 7 million container images stored in the Docker hub alone. In addition, the increase in private registries has also increased the performance and capacity requirements of the image storage system. Because there are a large number of redundant data and redundant files in the image file, using its high redundancy to reduce duplicate data can greatly reduce the ever-increasing registry storage requirements. Based on the characteristics of virtual machine images and container images, this paper proposes and designs a high-efficiency storage system for virtualized resource images given the large volume of virtualized resource images and high storage occupation. Through this system, the utilization rate of image storage can be improved and reduced the cost of store images.