A scheme for balancing heterogeneous request load in DHT-based P2P systems
Juan Li, Billy Ho Hung Cheung, Son T. Vuong · 2007
DHT-based P2P systems have been proven to be a scalable and efficient means of sharing information. With the entrance of quality of services concerns into DHT systems, however, the ability to guarantee that the system will not be overwhelmed due to load imbalance becomes much more significant, especially when factors such as item popularity and skewing are taken into consideration. In this paper, we focus on the problem of load imbalance caused by skewed access distribution. We propose an effective load balancing solution, which takes the peer heterogeneity and access popularity into account to determine the load distribution. Our algorithm achieves load balancing by dynamically balancing the query routing load and query answering load respectively. Experimentations performed over a Pastry-like system illustrate that our balancing algorithms effectively balance the system load and significantly improves performance.