Abstract: The main part of high speed network content auditing system(HSNCAS) is introduced. The process model of HSNCAS is classified into pipelining model and multi-section model. An algorithm of multi-priority memory feedback (MPMF) for pipelining model is presented and system software architecture is designed based on it. The MPMF algorithm forecasts and sends feedback of different priorities to the load balancer by the memory usage of every part of the system. It can effectively use the hardware resources and improve the overall data processing rate. A solution for “half session” is also proposed.

Key words: high speed network, content auditing, memory feedback, load balance

摘要： 介绍了高速网络内容审计系统的主要组成部分，提出了将内容审计系统的过程模型分为流水线模型和分段模型。提出了一种针对流水线模型的MPMF反馈算法，该算法利用系统各个部分的内存使用情况，对负载均衡器做出不同优先级的反馈，有效地利用了系统资源，并给出了一种满负荷情况下“半截会话”问题的解决方案。

关键词: 高速网络, 内容审计, 内存反馈, 负载均衡

CLC Number: 

• TP393