Dynamic load balancing in distributed virtual environments using heat diffusion

Abstract

Distributed virtual environments (DVEs) are attracting a lot of attention in recent years, due to the increasing popularity of online gaming and social networks. As the number of concurrent users of a DVE increases, a critical problem is on how the workload among multiple servers can be balanced in order to maintain real-time performance. Although a number of load balancing methods have been proposed, they either try to produce high quality load balancing results and become too slow or emphasize on efficiency and the load balancing results become less effective. In this article, we propose a new approach to address this problem based on heat diffusion. Our work has two main contributions. First, we propose a local and a global load balancing methods for DVEs based on heat diffusion. Second, we investigate two performance factors of the proposed methods, the convergence threshold and the load balancing interval. We have conducted a number of experiments to extensively evaluate the performance of the proposed methods. Our experimental results show that the proposed methods outperform existing methods in that our methods are effective in reducing server overloading while at the same time being efficient.

Dynamic Resource Management in Cloud-based Distributed Virtual Environments

“…This should be taken into account in load balancing, that is, the workload assigned to each server should be proportional to its capacity. However, most existing DVE dynamic load balancing techniques (such as [3] and references therein) do assume that servers are homogeneous. Hence, we extend the method proposed by [3] to handle the server heterogeneity when doing load balancing, i.e., the time when locally under-provisioned event arises but globally under-provisioned event doesn't as illustrated in Figure 1.…”

“…However, most existing DVE dynamic load balancing techniques (such as [3] and references therein) do assume that servers are homogeneous. Hence, we extend the method proposed by [3] to handle the server heterogeneity when doing load balancing, i.e., the time when locally under-provisioned event arises but globally under-provisioned event doesn't as illustrated in Figure 1. So far, we have presented our proposed resource management approach and the associated algorithms.…”

“…Upon spending up the chosen time period, he will move to a new region randomly chosen from the adjacent regions of his current region or a random remote region (in case for teleporting). The remote region one of the popular regions, called hotspots, which drives uneven user distributions and we set 20% of regions as hotspots (similar to [3]). Generally, users are moving faster and more frequently for exploration or questing than social interaction.…”

“…We compare these methods: BP, representing the bin packing based consolidation method [6]; NS, representing the method of dynamic load balancing with no scaling [3]; HS, representing our horizontal scaling based method; VS, representing our vertical scaling based method.…”

“…Specifically, we consider both global workload changes induced by user population fluctuations and local workload changes induced by user mobility. Our proposed approach not only improves resource efficiency as compared to those which overlook global workload dynamics (e.g., [3]), but also reduces service interruption induced by resource reallocation as compared to those which overlook local workload dynamics (e.g., [6]). We present the proposed approach and simulation results in Sections 2 and 3, respectively.…”