计算机应用与软件2016，Vol.33Issue(9)：27-32,6.DOI:10.3969/j.issn.1000-386x.2016.09.007
在线云存储服务的流量管理研究
ON TRAFFIC MANAGEMENT OF ONLINE CLOUD STORAGE SERVICE
摘要
Abstract
Cloud storage is an important branch of cloud computing applications,for effectively making use of bandwidth resources in data centre,it is important to design the bandwidth management with efficiency,equilibrium and good scalability and the traffic load balancing algo-rithm.Under the architecture of Dropbox which is the typical application of cloud storage service,it is able to realise load balance by designing the greedy algorithm with the priority of minimum bandwidth and the secondary random selection algorithm,and they are integrated with traffic predicting and bandwidth reservation techniques to realise a set of traffic load balancing and bandwidth reservation schemes.The greedy algo-rithm-based load balancing technique can archive good performance but is highly complex,communication costing and poor in scalability.The secondary random selection algorithm has low complexity while significantly decreases system communication cost.We test the two algorithms in experiments on both the real Dropbox traffic data and the large-scale simulated data,results show that the secondary random selection algo-rithm is able to achieve the balanced load scheduling in performance close to that obtained by the greedy algorithm.The traffic prediction-based bandwidth reservation technique ensures the QoS and raises the utilisation of network resource.关键词
云存储/负载均衡/流量管理/流量预测/带宽预留Key words
Cloud storage/Load balancing/Traffic management/Traffic prediction/Bandwidth reservation分类
信息技术与安全科学引用本文复制引用
李梦寒,郑小盈,李明齐,张宏鑫..在线云存储服务的流量管理研究[J].计算机应用与软件,2016,33(9):27-32,6.基金项目
国家自然科学基金项目（61100238）；中科院先导项目（XDA06010301）；中国科学院重点部署项目（KGZD －EW－103）；上海市科委项目（14510722300，13DZ1511200）；中国科学院青年创新促进会，浙江大学 CAD＆CG 国家重点实验室开放课题（A1314）。李梦寒，硕士生，主研领域通信网络。 （）
