摘要
提出了一种基于查询事件的日志模型,采用查询/应答日志匹配的方法完整的记录了一次查询事件,利用内存数据结构提高了海量数据写入的I/O效率;在日志分析过程中对日志文件建立二维哈希索引,利用布隆过滤器减少磁盘I/O次数,提高了分析效率.
In this paper, we investigate a logging model based on a complete DNS query event. The corresponding query/ response matching technique is designed, equipped with a facility of memory caching to solve the performance bottleneck of the insertion operations. Furthermore, a mechanism of two - dimensional hash indexing with Bloom Filter is designed to improve searching and retrieving by reducing disk I/O.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第2期95-98,102,共5页
Microelectronics & Computer
基金
中国科学院计算机网络信息中心青年基金项目(0814081105)
