Architecture and scalability of a high-speed traffic measurement platform with a highly flexible packet classification

Evolving network technologies, new web services and changing usage patterns continuously change traffic characteristics. But a thorough understanding of the traffic is the basis for many applications in networking. Thus, it is crucial to analyze up-to-date traffic traces collected by passive measurements in many relevant network contexts. As the traces’ quality defines the significance of such analysis, the measured data is required to be complete, temporally accurate and reliable. This is especially challenging for measurements on high-speed links.In this paper, we present a scalable architecture for a high-speed, passive measurement platform for obtaining highly accurate packet traces. Its functional blocks are distributed to a specialized hardware unit and a commodity PC unit according to their specific requirements. For pre-processing, the hardware unit integrates a protocol-aware classification and filtering module which allows an easy definition of classification and filtering rules. We analyze the platform’s scalability and support this by a implementation with current FPGA technology and a standard server PC.