No abstract

We propose a novel general-purpose network traffic Monitoring Application Programming Interface (MAPI) for network monitoring applications. Our work builds on a generalized network flow model that we argue is flexible enough to capture emerging application needs, and expressive enough to allow the system to exploit specialized monitoring hardware, where available. We describe an implementation of MAPI using the DAG 4.2 Gigabit Ethernet monitoring card and a commodity Gigabit Ethernet adapter, we present a set of experiments measuring overheads, and we demonstrate potential applications. Our experimental results suggest that MAPI has more expressive power than competing approaches, while at the same time is able to achieve significant performance improvements.

Abstract-Network monitoring and measurement is commonly regarded as an essential function for understanding, managing and improving the performance and security of network infrastructures. Traditional passive network monitoring approaches are not adequate for fine-grained performance measurements nor for security applications. In addition, many applications would benefit from monitoring data gathered at multiple vantage points within a network infrastructure.This paper presents the design and implementation of DiMAPI, an application programming interface for distributed passive network monitoring. DiMAPI extends the notion of the network flow with the scope attribute, which enables flow creation and manipulation over a set of local and remote monitoring sensors. Experiments with a number of applications on top of DiMAPI show that it has reasonable performance, while the response latency is very close to the actual round trip time between the monitoring application and the monitoring sensors. A broad range of monitoring applications can benefit from DiMAPI to efficiently perform advanced monitoring tasks over a potentially large number of passive monitoring sensors.

No abstract

Abstract. Effective network monitoring is vital for the growing number of control and management applications typically found in present-day networks. Increasing link speeds and the diversity of monitoring applications' needs have exposed severe limitations of existing monitoring techniques. As a response, the EU IST SCAMPI project designs and implements a scalable and programmable architecture for monitoring multi-gigabit networks. The SCAMPI architecture has an expressive programming interface, uses intelligent hardware, provides user policy management and resource control, and achieves scalability through parallelism. This paper addresses the problems with current high-speed network monitoring and presents the system architecture and components of the SCAMPI platform.