摘要
降低网络系统能耗是当前计算机领域构建绿色网络迫切需要解决的关键问题.网络设备的传输服务速率自适应性为优化网络能耗和提高网络能效提供了有效的途径,文中提出一种基于速率自适应的全局性和分布式的能耗优化路由策略.该策略从网络全局路由的角度出发,根据网络系统的服务特征,将为数据提供传输服务的网络组件抽象为一个处理域.为求解处理域中速率自适应时的服务速率和工作状态平均转换次数,把处理域的服务过程视为可变服务速率的服务系统.然后以网络系统总能耗最小化为目标,并满足相关的路由和性能等约束条件,建立基于速率自适应的网络能耗优化路由模型,利用改进的蚁群算法对模型进行求解.在仿真实验中,将文中提出的能耗优化路由的分布式启发算法与相关文献的OSPF和GreenOSPF节能路由算法进行比较,给出算法在能耗和延时方面的对比结果.多种实验情况下的对比结果表明,文中提出的能耗优化路由策略能更有效地匹配速率自适应机制,具有较好的节能效果,从而达到优化和降低能耗的目的.
Reducing network system energy consumption is important to construct green network in current computing area.Rate adaptation of network device provides an effective approach to optimize network energy consumption and improve network energy efficiency.In this paper,we propose a global and distributed routing strategy to optimize energy consumption based on rate adaptation.According to the characteristics of service in a network system,we establish an energy consumption optimization model by abstracting network service components as a processing domain from a global point of view and minimizing the total energy consumption of network system with related constraints.In order to obtain the rate of service and the average switched number of running mode in a processing domain,the service process of the processing domain is regarded as the service system with variable service rate.Then,we give an improvement ant colony optimization（ACO）algorithm to solve energy consumption optimization model.We obtain the comparison results of energy consumption and delay in NS-2 simulator.Compare with the OSPF andGreenOSPF energy efficiency algorithms in simulation,the energy consumption optimization routing strategy proposed by this paper is suitable for rate adaptation mechanism,and then obtains the goal of optimization and reducing of energy consumption.
出处
《计算机学报》
EI
CSCD
北大核心
2015年第3期555-566,共12页
Chinese Journal of Computers
基金
国家自然科学基金(61262003)
广西自然科学杰出青年基金(2013GXNSFGA019006)
广东省教育厅科技创新项目(2013KJCX0018)资助~~
关键词
网络系统
速率自适应
能耗优化
蚁群算法
network system
rate adaption
energy consumption optimization
ant colony algorithm
