Behavior of VNC in high-latency environments and techniques for improvement.

摘要

Thin-client computing offers many advantages over traditional PC computing including lower costs, ubiquitous access to resources, higher security, and easier maintenance. Unfortunately thin clients rely upon the services of other networked computers to properly function. Because of this tight-knit relationship, high-latency environments can render thin clients practically unusable. With the VNC thin-client system, performance can be capped at 1 frame per round-trip time (RTT) due to its client pull communication style. We extend VNC to build VNC-HL and show how to improve frame rate performance by employing PR, a pre-requesting technique, to periodically request updates, even with several previous requests pending. Experimental results demonstrate up to an order of magnitude of improvement. VNC-HL achieved 14 FPS of multimedia computing in a 500 ms latency network with negligible additional resource consumption. In this thesis we demonstrate that VNC can be improved to sustain high frame rate performance in high-latency environments by employing a periodic pre-request at the desired FPS rate.