Affiliations: [a] Dipartimento di Elettronica per l'Automazione, Università di Brescia, 25123 Brescia, Italy. E-mail: decapita@ing.unibs.it | [b] Computer Science Laboratory, SRI International, Menlo Park, CA 94025, USA. E-mail: lincoln@csl.sri.com | [c] Metanetworks Inc., 647 N. Santa Cruz Ave, Los Gatos, CA 95032, USA. E-mail: livio@metanetworks.org | [d] Dipartimento di Tecnologie dell'Informazione, Università di Milano, 20163 Crema, Italy. E-mail: samarati@dsi.unimi.it
Note: [1] A preliminary version of this paper appeared under the title “PGRIP: PNNI Global Routing Infrastructure Protection” in The Internet Society 1999 Network and Distributed System Security Symposium (NDSS’99), San Diego, CA, USA, February 1999.
Abstract: The development of new classes of distributed applications such as telephony, remote video, and virtual reality introduces new quality requirements that make inadequate even the current best-effort service model provided by international networks. Recent research activity has proposed new approaches to accommodate these new application requirements. While exploiting the resource needs of different applications to use the network resources efficiently, these approaches require the maintenance of a considerable amount of state information at the network nodes. Correctness and availability of such information are the basic requirements for the proper working of the network. In this paper we present a system, called Global Infrastructure Protection System (GIPS), to control improper modifications to this state information. We illustrate the GIPS's architecture and identify topology conditions to guarantee the distributed fault tolerant detection of anomalies. The system is based on the use of a hierarchical structure to organize and maintain the information at each node. Improper network states are described through rules that characterize state information updates that may result anomalous (or uncommon) with respect to the network status, past events occurred, or statistical measures. We introduce a notation to represent state information coming from heterogeneous protocols, and statistical operators to examine the history of state updates accumulated during operation. Finally, we present some examples using our notation to express heuristical rules detecting anomalous operations in a network.
