Analysis And Improvement Of Video Transmission In SPICE

In recent years, with the development of computer and network hardware and software technology, cloud computing is more and more becoming an important technology. As an important technical to support cloud computing, virtualization technology has become a focus in the industry. Actually as a typical technology for physical and logical isolation, virtualization technology is applied in many fields, such as operating systems and compiler theories. In recent years because it is deployed in lots of server-client systems, the virtualization technology is becoming a very active subject.Virtualization technology can enlarge the capacity of the hardware and simplify the software re-configuration process. CPU virtualization technology can simulate multi-CPU parallel computing on single CPU, which allows us to run multiple operating systems and program independently. So the efficiency of computer is improved significantly. SPICE is the protocol which is developed for Red Hat virtual desktop. It is an open source software, so new client modules or sub-extensions, are developed more easily.However in order to be deployed in client applications, SPICE contains some simple and unsatisfactory places. This thesis focuses on part of these problems. By strengthening the weak parts of SPICE, it will help domestic companies to enhance its core competitiveness of high-tech fields.This thesis discusses the background and problems of the SPICE protocol development, introduces the latest research status of virtualization protocols, and analyzes the inherent characteristics of the SPICE protocol and its video transmission and compression mechanisms. After comparison with other protocols, we propose SPICE video compression and transmission improvement goals.After the analysis of video compression transmission interface and system architecture, we carry out improvements. After comparing several basis functions, we selecte Haar wavelet function to replace discrete cosine transform. Then based on the features of SPICE video features and Haar wavelet function, the Huffman table used in entropy coding is updated, which make it more responsive to new compression algorithms. Meanwhile, we also improve the network adaption of video transmission with a more flexible and effective mechanisms. Exepriments indicate that the new SPICE protocol has been greatly improved, and the performance is better than Microsoft’s RDP7protocol.