Processing of Flow Accounting Data in Java: Framework Design and Performance Evaluation

Flow Accounting is a passive monitoring concept implemented in routers that gives insight into tra c behavior and network characteristics. However, processing of Flow Accounting data is a challenging task, especially in large networks where the rate of ow records received at the collector can be very high. We developed a tool for processing of Flow Accounting data written in Java. It provides processing blocks for aggregation, sorting, statistic, correlation and other tasks. Besides reading data from les for o ine analysis, it can also process data directly received from the network. In terms of multithreading and data handling, the tool is highly con gurable in order to optimize its performance for a given task. For setting these parameters there are several trade-o s concerning memory consumption and processing overhead. In this paper, we study these trade-o s based on a reference scenario and examine characteristics caused by garbage collection.