Cloud Storage Product Optimization Based On Swift And Cloud Virtual Machine Scheduling Algorithm Research

With the development of cloud computing, the demand for storage capacity and all areas of computing power provided by cloud computing are increasing. This paper mainly from the two angles researches the relevant product design and performance optimization solutions.Swift as an open source project Openstack’s storage component has been increasingly used and deployed because of its excellent stability and security, but it still can not be directly commercialized with many deficiencies, such as it can not support folder nesting mechanism and Rest user-oriented interface does not support CDMI (Cloud Data Management Interface) specification, also it does not support the realization of specific business scenarios. In this paper, based on Swift’s cloud storage service solutions, the paper proposed an improved service framework, the service framework implemented based on CMDI of Rest interface specifications can be used by different product systems. The framework can facilitate service module scalability, for example, it can easily add permissions management module and billing modules on this framework. This framework can abstract the underlying storage services, allowing users to easily use a unified standard Restful interface based on CDMI to call the underlying service, without the need for specific ISP and implementation details. This paper also explores the Swift poor performance solution in a small file storage and use Flashcache caching framework to some extent solve the small files stored solve performance problems.In cloud computing, cloud virtual machine tasks automatic scheduling (Auto-Scheduling) and cloud virtual machines automatically Scaling (Auto-Scaling) algorithm has been more research supported, how to automatically expand automated task scheduling and virtual machines to make research program running in the cloud computing performance of the virtual machine optimized got more attention, based on HEFT (Heterogeneous-Earliest-Finish-Time) algorithm for the corresponding cloud virtual machine automatic scheduling and automatic expansion algorithm is studied algorithm in a virtual machine scheduling priority budgetary issues, explores how without exceeding the established budget, to explore the virtual machine performance optimization solutions, and ultimately gives the experimental demonstration.