| In recent years,cloud computing has developed rapidly,and virtualization technology,one of the key technologies of cloud computing,has also developed rapidly.Virtualization technology has evolved from the early hardware-based virtualization to the current software-based virtualization,which also gave birth to the concept of desktop virtualization.As a mainstream desktop virtualization solution,the virtual desktop infrastructure emphasizes placing the user's desktop computing environment in the data center,and user only needs to access the remote virtual desktop through the network.Such a solution makes full use of the resources on the server side and solves the problem of excess resources on the client side.All the data between the server and the client are transmitted through the network,and the virtual desktop protocol is one of the key technologies.The transmission efficiency and processing method of the virtual desktop protocol directly determine the user experience.Currently,the mainstream commercial virtual desktop protocols on the market include RDP,ICA,PCoIP,and so on.In open source software,the SPICE is one of the few virtual desktop protocols that can be compared with commercial protocols.As an excellent open source virtual desktop protocol,the SPICE is basically complete and can work well in most intranet environments,but there are still many optimization spaces in some specific cloud computing environments.This thesis analyzes the mainstream commercial and open source virtual desktop protocols in the market in a specific cloud computing environment.It was found that some of these protocols have certain use restrictions,such as the inability to support high-definition video playback,large 3D animation processing in a WAN environment,data transmission occupying too much bandwidth,and certain performance requirements for terminal devices.Under this background,this thesis conducts an in-depth analysis of the open source protocol SPICE,and deeply understands the design core and key technologies of the SPICE.Then,in view of the shortcomings in a specific cloud computing environment,this thesis proposes a corresponding optimization scheme.Firstly,this thesis proposes to construct a network detection channel based on the SPICE protocol's virtual multi-channel technology.The network environment in the SPICE connection process is detected in real time,and the detection results are analyzed and evaluated.Then adjust the transmission priorities and image processing methods of different virtual channels to meet different network environments.Secondly,in view of the problem that the SPICE protocol is not optimized for video data processing,this thesis uses the more efficient video coding technology HEVC to replace the coding algorithm in the original protocol.In the design of this thesis,corresponding optimization schemes are also proposed for special scenarios such as high interaction.Finally,according to the actual needs in the cloud platform,such as USB device redirection,client internationalization,etc.,SPICE is functionally improved.The purpose of this thesis for SPIEC optimization is to solve the problem of poor performance of existing virtual desktop protocols in specific cloud computing environments and cloud platforms.On the basis of SPICE's original protocol,the existing protocol is optimized with high availability,and high scalability.In the end,the optimized protocol has approximately doubled its response speed in high-interaction scenarios.The amount of bandwidth occupied during video playback is reduced to about 12% of the original. |
PDF Full Text Request