Abstract
Second Life is a multi-purpose online virtual world that is increasingly being used for applications and simulations in diversified areas such as education, training, entertainment, and even for applications related to Artificial Intelligence. For the successful implementation and analysis of most of these applications, it is important to have a robust mechanism to extract low-level data from Second Life in high frequency and high accuracy. However, currently Second Life does not have a reliable or scalable inbuilt data extraction mechanism, nor the related research provides a better alternative. This paper presents a robust and reliable data extraction mechanism from Second Life. We also investigate the currently existing data extraction mechanisms in detail, identifying their limitations in extracting data with high accuracy and high frequency.