[show abstract][hide abstract] ABSTRACT: We propose an approach to remote display systems in which the client predicts the screen update events that the server will send and applies them to the screen im- mediately, thus eliminating the network round-trip time and making the system more responsive in a wide-area or high loss environment. Incorrectly predicted events are undone when the actual events arrive from the server. The approach requires no server or protocol changes, and thus can work with existing systems. Since it is core to the feasibility of such a speculative remote display sys- tem, we study the predictability of the events that occur under typical workloads in two extant systems, Windows Remote Desktop and VNC. We find that simple, state- limited Markov models are often able to correctly predict the next event. Based on these results, we design, imple- ment, and evaluate a speculative remote display exten- sion to the VNC client. In our implementation, the end user can trade off between the responsiveness of the dis- play and the level of temporarily displayed incorrect pre- dictions. We evaluate VNC/SRD with two user studies. We conclude by describing design alternatives.
2008 USENIX Annual Technical Conference, Boston, MA, USA, June 22-27, 2008. Proceedings; 01/2008
[show abstract][hide abstract] ABSTRACT: Experimental computer systems research typically ignores the end-user, modeling him, if at all, in overly simple ways. We argue that this (1) results in inadequate performance evaluation of the systems, and (2) ignores opportunities. We summarize our experiences with (a) directly evaluating user satisfaction and (b) incorporating user feedback in dieren t areas of client/server computing, and use our experiences to motivate principles for that domain. Specically , we report on user studies to measure user satisfaction with resource borrowing and with dieren t clock frequencies in desktop computing, the development and evaluation of user inter- faces to integrate user feedback into scheduling and clock frequency decisions in this context, and results in predicting user action and system response in a remote display system. We also present initial results on extending our work to user control of scheduling and mapping of virtual machines in a virtualization-based distributed computing environment. We then generalize (a) and (b) as recommendations for in- corporating the user into experimental computer systems research.
Proceedings of the Workshop on Experimental Computer Science, Part of ACM FCRC, San Diego, 13-14 June 2007; 01/2007