Internet routers generally see packets from a fast flow more often than a slow flow. This suggests that network fairness may be improved without per-flow information. In this paper, we propose a scheme using Most Rece...Internet routers generally see packets from a fast flow more often than a slow flow. This suggests that network fairness may be improved without per-flow information. In this paper, we propose a scheme using Most Recently Used List (MRUL)-a list storing statistics of limited active flows that sorted in most recently seen first mode-to improve the fairness of RED. Based on the list, our proposed scheme jointly considers the identification and punish of the fast and unresponsive fast flows, and the protection of slow flows. Its performance improvements are demonstrated with extensive simulations. Different from the previous proposals, the complexity of our proposed scheme is proportional to the size of the MRUL list but not coupled with the queue buffer size or the number of active flows, so it is scalable and suitable for various routers. In addition, another issue we address in this paper is queue management in RED. Specifically, we replace the linear packet dropping function in RED by a judicially designed nonlinear quadratic function, while original RED remains unchanged. We call this new scheme Nonlinear RED, or NLRED. The underlying idea is that, with the proposed nonlinear packet dropping function, packet dropping becomes gentler than RED at light traffic load but more aggressive at heavy load. As a result, at light traffic load, NLRED encourages the router to operate in a range of average queue sizes rather than a fixed one. When the load is heavy and the average queue size approaches the pre-determined maximum threshold (i.e. the queue size may soon get out of control), NLRED allows more aggressive packet dropping to back off from it. Simulations demonstrate that NLRED achieves a higher and more stable throughput than RED and REM. Since NLRED is fully compatible with RED, we can easily upgrade/replace the existing RED implementations by NLRED.展开更多
In this paper, we propose a new mechanism called explicit rate notification(ERN) to be used in end-to-end communications. The ERN scheme encodes in the header of transmission control protocol(TCP) packets information ...In this paper, we propose a new mechanism called explicit rate notification(ERN) to be used in end-to-end communications. The ERN scheme encodes in the header of transmission control protocol(TCP) packets information about the sending rate and the round trip time(RTT) of the flows. This new available information to the intermediate nodes(routers) is used to improve fairness, increase utilization, decrease the number of drops, and minimize queueing delays. Thus, it induces a better management of the queue. A comparison of our scheme with preexistent schemes, like the explicit congestion notification scheme, shows the effectiveness of the proposed mechanism.展开更多
Improving the Quality of Service (QoS) of Internet traffic is widely recognized as a critical issue for the next-generation networks. In this paper, we present a new algorithm for the active queue management, namely R...Improving the Quality of Service (QoS) of Internet traffic is widely recognized as a critical issue for the next-generation networks. In this paper, we present a new algorithm for the active queue management, namely RED-DTB. This buffer control technique is used to enforce approximate fairness among a large number of concurrent Internet flows. Like RED (Random Early Detection) algorithm, the RED-DTB mechanism can be deployed to actively respond to the gateway congestion, keep the gateway in a healthy state, and protect the fragile flows from being stolen bandwidth by greedy ones. The algorithm is based on the so-called Dual Token Bucket (DTB) pattern. That is, on the one hand, every flow is rate-limited by its own token bucket, to ensure that it can not consume more than its fair share of bandwidth; On the other hand, to make some compensations to less aggressive flows, such as connections with larger round trip time or smaller sending window, and to gain a relatively higher system utilization coefficient, all flows, depending on their individual behavior, may have a chance to fetch tokens from the public token bucket when they run out of their own share of tokens. The algorithm is analyzed and evaluated by simulations, and is proved to be effective in protecting the gateway buffer and controlling the fair allocation of bandwidth among flows.展开更多
An efficient congestion control approach should gain weighted fairness and also robustness against external disturbances.Moreover,variations in the number of active sources cause a network to operate as a switched sys...An efficient congestion control approach should gain weighted fairness and also robustness against external disturbances.Moreover,variations in the number of active sources cause a network to operate as a switched system in which the initial values affect the transient response.The main contribution of this article is to design a rate-based adaptive robust queue management contemplating the initial value compensation(IVC).The congestion is avoided via Two-Degree-of-Freedom Internal Model Control protocol as a robust procedure.The utilisation factors are designated according to the delay of each source(weighted fairness)in which the sources will use the maximum value of the bottleneck capacity(maximum utilisation).On the other hand,the number of the users is determined adaptively and effects of non-zero initial values are eliminated through a protocol called IVC.Simulation results through Network Simulator 2 and Simulink software confirm the analytical results and the efficacy of the proposed method.展开更多
A new weighted fair queueing algorithm is proposed, which uses the novel flow-based service ratio parameters to schedule flows. This solves the main drawback of traditional weighted fair queneing algorithms- the packe...A new weighted fair queueing algorithm is proposed, which uses the novel flow-based service ratio parameters to schedule flows. This solves the main drawback of traditional weighted fair queneing algorithms- the packet-based calculation of the weight parameters. In addition, this paper proposes a novel service ratio calculation method and a queue mangement technology. The former adjusts the service ratio parameters adaptively based on the dynamics of the packet lengths and thee solves the unfairness problem induced by the variable packet length. The latter improves the utilization of the server＇s queue buffer and reduces the delay jitter through restricting the buffer length for each flow.展开更多
Based on the proportionally fair scheme that Kelly proposed to solve the optimization problems for utility function in networks, and in order to improve the congestion control performance for the queue in router, the ...Based on the proportionally fair scheme that Kelly proposed to solve the optimization problems for utility function in networks, and in order to improve the congestion control performance for the queue in router, the linear and terminal sliding active queue management (AQM) algorithms are designed. Especially in the ter-minal sliding AQM algorithm, a special nonlinear terminal sliding surface is designed in order to force queue length to reach the desired value in finite time. The upper bound of the time is also obtained. Simulation re-sults demonstrate that the proposed congestion algorithm enables the system be better transient and stable performance. At the same time, the robustness is guaranteed.展开更多
针对云计算系统在运行过程中由于计算节点空闲而产生大量空闲能耗,以及由于不匹配任务调度而产生大量"奢侈"能耗的能耗浪费问题,提出一种通过任务调度方式的能耗优化管理方法.首先,用排队模型对云计算系统进行建模,分析云计...针对云计算系统在运行过程中由于计算节点空闲而产生大量空闲能耗,以及由于不匹配任务调度而产生大量"奢侈"能耗的能耗浪费问题,提出一种通过任务调度方式的能耗优化管理方法.首先,用排队模型对云计算系统进行建模,分析云计算系统的平均响应时间和平均功率,建立云计算系统的能耗模型.然后提出基于大服务强度和小执行能耗的任务调度策略,分别针对空闲能耗和"奢侈"能耗进行优化控制.基于该调度策略,设计满足性能约束的最小期望执行能耗调度算法ME3PC(minimum expectation execution energy with performance constraints).实验结果表明,该算法在保证执行性能的前提下,可大幅度降低云计算系统的能耗开销.展开更多
文摘Internet routers generally see packets from a fast flow more often than a slow flow. This suggests that network fairness may be improved without per-flow information. In this paper, we propose a scheme using Most Recently Used List (MRUL)-a list storing statistics of limited active flows that sorted in most recently seen first mode-to improve the fairness of RED. Based on the list, our proposed scheme jointly considers the identification and punish of the fast and unresponsive fast flows, and the protection of slow flows. Its performance improvements are demonstrated with extensive simulations. Different from the previous proposals, the complexity of our proposed scheme is proportional to the size of the MRUL list but not coupled with the queue buffer size or the number of active flows, so it is scalable and suitable for various routers. In addition, another issue we address in this paper is queue management in RED. Specifically, we replace the linear packet dropping function in RED by a judicially designed nonlinear quadratic function, while original RED remains unchanged. We call this new scheme Nonlinear RED, or NLRED. The underlying idea is that, with the proposed nonlinear packet dropping function, packet dropping becomes gentler than RED at light traffic load but more aggressive at heavy load. As a result, at light traffic load, NLRED encourages the router to operate in a range of average queue sizes rather than a fixed one. When the load is heavy and the average queue size approaches the pre-determined maximum threshold (i.e. the queue size may soon get out of control), NLRED allows more aggressive packet dropping to back off from it. Simulations demonstrate that NLRED achieves a higher and more stable throughput than RED and REM. Since NLRED is fully compatible with RED, we can easily upgrade/replace the existing RED implementations by NLRED.
文摘In this paper, we propose a new mechanism called explicit rate notification(ERN) to be used in end-to-end communications. The ERN scheme encodes in the header of transmission control protocol(TCP) packets information about the sending rate and the round trip time(RTT) of the flows. This new available information to the intermediate nodes(routers) is used to improve fairness, increase utilization, decrease the number of drops, and minimize queueing delays. Thus, it induces a better management of the queue. A comparison of our scheme with preexistent schemes, like the explicit congestion notification scheme, shows the effectiveness of the proposed mechanism.
基金the National Natural Science Foundation of China(60132030)and the National Education Department Doctorial Foundation Project(RFDP1999048602)
文摘Improving the Quality of Service (QoS) of Internet traffic is widely recognized as a critical issue for the next-generation networks. In this paper, we present a new algorithm for the active queue management, namely RED-DTB. This buffer control technique is used to enforce approximate fairness among a large number of concurrent Internet flows. Like RED (Random Early Detection) algorithm, the RED-DTB mechanism can be deployed to actively respond to the gateway congestion, keep the gateway in a healthy state, and protect the fragile flows from being stolen bandwidth by greedy ones. The algorithm is based on the so-called Dual Token Bucket (DTB) pattern. That is, on the one hand, every flow is rate-limited by its own token bucket, to ensure that it can not consume more than its fair share of bandwidth; On the other hand, to make some compensations to less aggressive flows, such as connections with larger round trip time or smaller sending window, and to gain a relatively higher system utilization coefficient, all flows, depending on their individual behavior, may have a chance to fetch tokens from the public token bucket when they run out of their own share of tokens. The algorithm is analyzed and evaluated by simulations, and is proved to be effective in protecting the gateway buffer and controlling the fair allocation of bandwidth among flows.
文摘An efficient congestion control approach should gain weighted fairness and also robustness against external disturbances.Moreover,variations in the number of active sources cause a network to operate as a switched system in which the initial values affect the transient response.The main contribution of this article is to design a rate-based adaptive robust queue management contemplating the initial value compensation(IVC).The congestion is avoided via Two-Degree-of-Freedom Internal Model Control protocol as a robust procedure.The utilisation factors are designated according to the delay of each source(weighted fairness)in which the sources will use the maximum value of the bottleneck capacity(maximum utilisation).On the other hand,the number of the users is determined adaptively and effects of non-zero initial values are eliminated through a protocol called IVC.Simulation results through Network Simulator 2 and Simulink software confirm the analytical results and the efficacy of the proposed method.
基金National Natural Science Foundation of China ( No.60572157)Sharp Corporation of Japanthe Hi-Tech Research and Development Program(863) of China (No.2003AA123310)
文摘A new weighted fair queueing algorithm is proposed, which uses the novel flow-based service ratio parameters to schedule flows. This solves the main drawback of traditional weighted fair queneing algorithms- the packet-based calculation of the weight parameters. In addition, this paper proposes a novel service ratio calculation method and a queue mangement technology. The former adjusts the service ratio parameters adaptively based on the dynamics of the packet lengths and thee solves the unfairness problem induced by the variable packet length. The latter improves the utilization of the server＇s queue buffer and reduces the delay jitter through restricting the buffer length for each flow.
文摘Based on the proportionally fair scheme that Kelly proposed to solve the optimization problems for utility function in networks, and in order to improve the congestion control performance for the queue in router, the linear and terminal sliding active queue management (AQM) algorithms are designed. Especially in the ter-minal sliding AQM algorithm, a special nonlinear terminal sliding surface is designed in order to force queue length to reach the desired value in finite time. The upper bound of the time is also obtained. Simulation re-sults demonstrate that the proposed congestion algorithm enables the system be better transient and stable performance. At the same time, the robustness is guaranteed.
文摘针对云计算系统在运行过程中由于计算节点空闲而产生大量空闲能耗,以及由于不匹配任务调度而产生大量"奢侈"能耗的能耗浪费问题,提出一种通过任务调度方式的能耗优化管理方法.首先,用排队模型对云计算系统进行建模,分析云计算系统的平均响应时间和平均功率,建立云计算系统的能耗模型.然后提出基于大服务强度和小执行能耗的任务调度策略,分别针对空闲能耗和"奢侈"能耗进行优化控制.基于该调度策略,设计满足性能约束的最小期望执行能耗调度算法ME3PC(minimum expectation execution energy with performance constraints).实验结果表明,该算法在保证执行性能的前提下,可大幅度降低云计算系统的能耗开销.
