作者: Tanya Bragin , Roxana Geambasu , Magdalena Balazinska , Jaeyeon Jung
DOI:
关键词:
摘要: Today, network intrusion detection systems (NIDSs) use custom solutions to log historical flows and support forensic analysis by administrators. These are expensive, inefficient, lack flexibility. In this paper, we investigate database for interactive analysis. We show that an "out-of-the-box" relational management system (RDBMS) can moderate flow rates in a manner ensures high query performance. To enable significantly higher data rates, propose technique based on on-demand view materialization indexing. our approach, when event occurs, the proactively extracts relevant indexes it preparation queries over data. approach improves response times large class of queries, while maintaining insert throughput.