A multi-server approach for large scale collaborative game-based learning

Abstract

E-learning through online games, where users play collaboratively to gain knowledge, has great potential to significantly change the way we learn. As the number of participants is no longer limited by the classroom, the learning process could potentially involve tens of thousands of learners. However, hosting massive users playing in a shared game world is nontrivial, as the underlying servers may get overloaded by the constantly changing workload due to user activity. In this work, we adapt a multi-server approach with dynamic load balancing to enable large scale collaborative game-based learning. Through simulation, we thoroughly evaluate its performance and identify the optimal settings of the key load balancing parameters under different scenarios. Results imply that the multi-server approach can support tens of thousands of users learning together, through combining the power of multiple servers each of which only can handle hundreds of users. © 2014 Springer International Publishing.