为了解决复杂任务群调度过程中资源利用不均、任务完成时间较长等问题,以最小化资源负载均方差和最小化任务群完成时间为目标构建复杂任务群资源调度模型,提出一种融合局部搜索和Pareto支配的多目标优化算法BRLSN(multi-objective optimization based on boundary range local search and NSGA-Ⅱ,BRLSN)。该算法采用有效的编码方式与交叉变异算子进行迭代寻优,并利用基于边界区域局部搜索的精英保留策略扩大算法搜索范围,保存种群优良个体。实验结果表明,BRLSN相较于其他多目标算法在收敛性和多样性上有显著的提升,同时算法收敛速度更快,种群质量更高,明显优化了最终目标函数的结果值。
To enhance user perceived performance,Akamai' s content distribution network(CDN) utilizes the domain name system(DNS) effectively to redirect users to close-by content replicas over short time scale.The use of DNS redirection has brought a significant advantage to Akamai' s CDN.However,the reliance on DNS also poses a fundamental threat to Akamai' s content distribution model.In particular,systematic evaluation and quantification illustrates the impact of recent emerging public DNS resolving services on Akamai' s CDN,including the degradation of corresponding user perceived performance and the benefit damnification of collaborating internet services providers(ISPs) that host Akamai's edge servers.The measurement demonstrates that a considerable penetration of public DNS resolving services(e.g.,OpenDNS and GoogleDNS) effectively corrupts the Akamai' s serving model,namely the large-scale server distribution and quick DNS redirection.
基于OFDMA-WLAN系统提出了一种改进的上行MAC协议,用以提高IEEE802.11无线局域网(Wireless Local Area Network,WLAN)在大量站点(stations,STAs)存在时的信道利用率和吞吐量。在OFDMA-WLAN系统存在多个子信道时,该协议使多个STA同时在时域与频域上按照CSMA/CA方式发送信道接入的请求帧,在不需要额外增加信道与射频的基础上,解决了随机接入带来的静默问题。数值分析结果表明,与现有的基于OFDMA的上行MAC协议FICA相比,本协议较好地提高了系统的信道利用率和吞吐量。