As an infrastructure for data distribution, overlay networks must incorporate efficient routing and adequate robustness in order to achieve fast and accurate data distribution in an environment with a high node churn.Considering that the existing overlay networks mostly focus on a single optimization objective and fail to ensure routing efficiency and robustness.Simultaneously, a hybrid overlay network for data distribution, Laurel, is proposed in this paper.Laurel achieves a better trade-off between routing efficiency and robustness by combining the inter-cluster multiple structured topologies with the intra-cluster unstructured topologies.Laurel also provides mechanisms for a dynamic, concurrent cluster creation, cluster departure, and load balance to make data distribution more adaptive to the dynamic network environment.Experimental results show that compared with existing overlay networks, Laurel can support faster and more accurate data distribution, even when a large amount of nodes fail in the system and balance the load within clusters.%覆盖网是各种数据分发应用的基础架构.在节点波动的网络环境中实现快速而准确的数据分发,对覆盖网提出了两个要求:高效的数据路由;较强的系统鲁棒性.已有的覆盖网构建方法多侧重于某个方面的优化,因而未能充分权衡数据路由效率与系统鲁棒性.提出了一种混合式数据分发覆盖网--Laurel.Laurel通过簇间多重结构化拓扑与簇内非结构化拓扑的结合,实现了路由效率与鲁棒性的高效折衷,并通过簇动态创建、退出以及负载平衡机制增强了对动态变化环境的适应能力.实验结果表明,相对于已有方法,Laurel即使在节点频繁波动的网络环境中也能快速而准确地分发数据,并且具有较好的负载平衡效果.
展开▼
