Towards Characterizing Second Life
报告人 : Prof. Wei Tsang Ooi (National University of Singapore)
In the first part of this talk, I will describe our effort in collecting mobility traces of avatars in Second Life, a large-scale, multi-user virtual world. I will present some of the findings from our traces that confirm common understanding of how avatars behave, as well as findings that yield new insights to their behavior. I will describe how these findings provide hints to design of a better multi-user virtual world.
In the second part of the talk, I will describe our preliminary effort in characterizing and caching of textures in Second Life. We found that, surprisingly, textures constitute up to 85% of all data in Second Life. This finding leads us to design and implement a caching proxy for textures for Second Life. Using the insights we gain from our traces, we design a caching algorithm for textures that improves the hit rate compared to using standard caching algorithms.