论文部分内容阅读
在网络非终端部分对数据流的处理在传统上就面对很多问题,例如数据回环、差错包、IP分片等。而现代网络随硬件水平的提高,数据量大幅度增加亦为网络流数据处理带来的巨大的难度。除此之外数据处理还要受到了内存、CPU、I/O速度的严重限制。为了通过软件实现有效的网络数据流重组,重点研究快速正确地对IP报文分析重组。研究集中在减少内存占用,提高处理速度,减少错误率,消灭错误包带来的不稳定性。
In the non-terminal part of the network to deal with the data flow traditionally faced many problems, such as data loopback, error packets, IP fragmentation. The modern network with the improvement of hardware level, a substantial increase in the amount of data for network streaming data processing also brings great difficulty. In addition to data processing but also by the memory, CPU, I / O speed severely limited. In order to realize effective network data flow reorganization through software, the focus is on analyzing and reassembling IP packets quickly and correctly. Research has focused on reducing memory footprint, increasing processing speed, reducing error rates, and eliminating the erratic nature of error packets.