Section: Contracts and Grants with Industry

INRIA-ALU joint lab

Participants : Humberto Abdelnur, Laurent Andrey, Rémi Badonnel, Olivier Festor [ Contact ] .

Dates

July 2008 - December 2011

Partners

Alcatel Lucent, INRIA.

This joint lab brings together research teams from INRIA and Alcatel Lucent Bell Labs for addressing the key challenges of autonomous networking in three critical areas: semantic networking, high manageability and self-organized networks. Our activity is part of the joint initiative dedicated to high manageability, and focuses on security management aspects with the Alcatel-Lucent Bell Labs teams on network security. Our work in this joint lab concerns the automation of security management. It includes a first activity related to fuzzing, which includes the improvement of the KiF framework as well as the design of novel fuzzing models for Alcatel-Lucent specific protocols. A second activity of the joint lab aims at investigating to what extent risk management strategies can be applied to VoIP infrastructures. The objective is to design and experiment dynamic risk management methods and techniques for voice oriented critical services.

In 2009, we have pursued our activity on fuzzing methods by completing the deployment testbed and by specifying more elaborated testing scenarios. In particular we have removed the SIP-specific features of our fuzzing tool so that we can specify testing scenarios with several different protocols. In the meantime we have developed new techniques allowing our fuzzing tool to generate a behaviour learnt from captured traces. Our efforts also focused on a runtime risk management strategy for preventing risks in VoIP networks and services. This strategy aims at dynamically altering the network exposure in a graduated manner in order to limit the impact of security safeguards on the VoIP service performances. Risk management provides new perspectives with respect to that issue. Risk is typically defined as the combination of the probability that a given threat exercises a vulnerability and the resulting impact of that adverse events on the network infrastructure. Risk management is the process consisting of identifying risks, assessing and evaluating them, and taking steps (security safeguards) to reduce risks to an acceptable level. In that context, we have extended a risk model, have specified a dedicated architecture and have evaluated the solution through a set of experiments in the case scenario of SPIT attacks. To ease the tracking of VoIP activity on heterogeneous environments, we have worked with Vijay Gurbani from Bell-labs on the design and early prototyping of a common log format for SIP entities. This format is now in its second release as an IETF draft [50] , [49] . We pursue the standardization activity on this item.