Towards Smarter Probes: In-Network Traffic Capturing and Processing

Abstract

Monitoring is a matter of the greatest importance for the correct operation of current communication networks. In spite of that, analyzing and checking out the traffic flowing over a high capacity link is still a very challenging technological issue, due to the huge amount of data stemming from such a process. Furthermore, current national and international legislation is imposing stricter and stricter limits on the storage and utilization of potentially privacy-sensitive data that may be generated from monitoring applications. We argue that both of these problems can be effectively addressed by increasing and extending the capabilities of traffic capturing devices beyond plain packet capturing and flow metering. Therefore, we envision a new generation of smart probes that support traffic pre-processing according to the needs of the specific application that is expected to provide the final results of the monitoring activity. The benefits of such an approach are two-fold: on one hand, in-network traffic filtering allows to discard a huge amount of information which is not relevant at all to the selected application, thus relaxing the performance requirements of the application itself. On the other hand, traffic pre-processing can be used to hide personal information that may be made available only to a user in possession of the required privileges upon verification of a given condition. Following such a general approach we propose a modular architecture that allows application specific traffic pre-processing to be carried out in a scalable and performance-effective way. Such an architecture interacts with the external network by enforcing strict role-based policies, thus allowing selective and proportional information disclosure; the architecture as it is can be easily integrated with a standard access control infrastructure. An example application is demonstrated in order to prove the effectiveness of the proposal.