The new PETERPAN Project: an IP-ATM integration point of view

Due to the explosion of multimedia, the traffic over Internet is growing not only as far as volume is concerned, but, even more important, it is changing in its nature; more and more new emerging applications are Quality of Service demanding; they need bounded transfer delay and reliability of the timely information delivery. The real time characteristic of multimedia applications requires a service architecture able to guarantee QoS, thus the traditional best effort paradigm is simply no longer sufficient. The networking community has witnessed an emerging breed of service differentiation protocols to provide some flows with better network service and quality.

"Within the telecommunication world, along with the explosion of Internet that is inventing structures for traffic handling to sustain the QoS as requested by the different applications, we witness the other reality: the ATM. It has been designed with intrinsic characteristics of scalability and reliability, with the objective of providing a common network infrastructure for broadband applications" says Giancarlo Rigolio, the PETERPAN project leader. As a matter of fact several PNOs have always believed in ATM, and today the commercial offering of ATM SVC is a reality. The ATM QoS is intrinsically in the ATM and we can offer with it a set of selected traffic categories with precise commitment. When Internet real time applications run over an ATM network, synergy can be identified between the Internet and the ATM. Sometimes the QoS (TCP, RSVP, ...) control mechanisms embedded into the higher layer and applied over ATM lead to a duplicated and often contending dynamics in the traffic control at the ATM layer. To avoid the consequent waste of resources, comprehensive network strategies have to be developed.

The evolution of Internet, perhaps in a long temporal perspective, may proceed in the direction of native ATM solutions. However, the migration (if any) from IP-based to native-ATM solution is not expected to be generalised, and perhaps mixed scenarios will last for long. Therefore, it is really hard to guess for the next future both the quota of ATM-native traffic and the IP-RSVP quota and their evolution. An unquestionable fact is that the future of Internet will depend on the capability of the network to efficiently handle the IP traffic in general and persistent flows in particular. Anyhow, the solution for IP handling shall be designed to be able to coexist and co-operate, in a fully integrated way, with solutions for pure ATM traffic support. This will be in a hybrid network architecture which spans over the access as well the core network elements.

"Integration of Internet and ATM is the issue considered by the PETERPAN Project as reported on the PETERTAN web site" continues Mr. Rigolio, "and it addresses efficient use of infrastructure through control of traffic flow and network resources under continuous consideration of QoS and user related parameters". Integrated solutions harmonising independent solutions of ATM and Internet are sought throughout, and the merits of Internet and ATM are exploited case by case.

The point of view of all the Project is to identify and develop an open architecture, in opposition to closed proprietary solutions, making full use of standard capabilities already offered by ATM in both control and user planes. The Project is not claiming the invention of a new service architecture, but the goal is to fully exploit standard features (as opposed to proprietary solutions) already made available by ATM and at the same time to support approaches currently discussed in IETF. The main objective of this action is to save investments already done by public network operators but also by end users and private networks or access network operators. It is expected that an architecture running efficiently IP applications over ATM will provide a very good solution to support the access of information; this will allow a still faster acceptance of the Internet concept by the general public.

The Project involves system and traffic studies as well as development, implementation, and testing of enhanced functionality for control and user plane. Field trials in several experimental islands will constitute the environment where the effectiveness of the solution identified will be proven. In addition, PETERPAN will perform accurate characterisation of IP over ATM traffic and the development of specific measurement and modeling techniques.

Given the intrinsic world-wide "scope" of Internet, the Project has deemed appropriate to assume the widest possible perspective, involving partners not only from Europe but also from overseas (USA), where furthermore one of the trial islands has been localised. This will allow the testing of the solution identified by the Project over a multi-vendor environment, with an intercontinental perspective. The consortium accounts for companies representative of the major players in the evolution of the world wide communication infrastructure, namely network operators, Telecom and IT manufacturers, and also relies on the co-operation of research centres and universities, traditionally leading the research on Internet issues.

In the framework of the ACTS Programme, the project, which refers to task AC321 "Convergence and Integration: Internet-ATM" is included in the Domain 3 (High Speed Networking) and is active in NI chain (Network Interworking ) group.

By Gianfranco Macalleno, CSELT, 12.05.1998

 News Index