欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
由于硬件、软件或传输故障等，用于流量矩阵估计的简单网络管理协议（Simple Network Mamagement Protocol，SNMP）数据可能包含脏数据，从而影响流量矩阵的精度．针对这个问题，提出一种基于SNMP的脏数据处理模型，摆脱了原有SNMP脏数据处理需要源-目的节点对间流量大规模测量的限制．基于交替投影方法，对此模型提出求得L0范数最小的稀疏脏数据处理方法．该算法降低了网络测量开销和时间复杂度，易于实现．实验表明，该算法对脏数据校正也有较高精度．
英文摘要：
Due to the hardware/software/transmission faults, the SNMP date used in traffic matrices estimating might contain dirty date and impact its accuracy. To solve this problem, a SNMP-self-based model, which was free for largely traffic measurement of Origin/Destination pairs in previous SNMP dirty data transaction, was pres- ented in this paper. Furthermore, based on alternative projection, a sparse dirty data processing algorithm was proposed to get L0 norm minimized Solution. It has lower measuring overhead and calculation complexity. Also, it is adaptive to applicationwidely. Besides, it can get high precision.
同期刊论文项目
同项目期刊论文
期刊信息
位置：