Reducing the Visibility of High Latency Transitions in a Virtual World

Online virtual worlds have gradually gained prominence in the age of the internet. While most of them exist as video games, some have aims of being a tool for general use; often using the term "metaverse" to describe themselves. Some of these worlds distribute their load across multiple servers, allowing users to transition between servers seamlessly. These severs, however, are typically local to each other, so latency between them is not an issue. This thesis makes the case for a seamless virtual world with servers spread across the globe, opening the door for, or even incentivizing cultural exchange. Three strategies, each building on the last, are proposed for reducing the the visibility of non-trivial latency between servers. While virtual worlds on multiple servers have existed for many years, there appears to be little to no public research on the subject of this thesis: inter-server latency. For this purpose, a virtual world was implemented; servers to run it and clients to interact with it. A delayer to artificially create latency between devices completed the core components needed for the research to proceed. The most advanced synchronization strategy implemented, dubbed Weighted Overlap, is designed to significantly reduce visible artifacts of high latency inter-server transfer. The results show that Weighted Overlap makes transitions appear significantly smoother to clients than both a baseline implementation and the two less advanced implemented strategies. This research demonstrates that, at least with some trade-offs, latency in the demanding context of seamless virtual worlds is not impossible to overcome