SGER: System-Level Network Modeling
Project Award Date: 09-01-2003
Modern communication networks are shifting from providing a basic data transport capability to providing a rich collection of network services. Recent advanced services envisioned by network designers include: multicast, reliable and adaptive multicast, resource reservation, caching services, and mobile access. Even these services and capabilities are eclipsed by needs for network protection, distributed web-page composition, repelling denial of service attacks, and assured network communications. The underlying network structure used to deliver these services is rapidly changing, becoming increasingly mobile and dynamic. We are embracing more wireless infrastructure, more mobile computing devices, increased interaction between services, and increased dynamic network services. To achieve desired capabilities in the increasingly dynamic network environment, dramatic changes in networking infrastructure are being introduced to provide a rich and highly dynamic networking environment. Yet, to the user, the network must be consistent and predictable. These changes pose new challenges in the way network systems and services are designed and deployed.
We will explore new techniques for design of advanced networking services. Specifically, we will concentrate on understanding and predicting emergent behaviors resulting from interactions between and among services and infrastructure elements. Our hypothesis is that using new, rigorous system-level design techniques, network infrastructure can be more flexible while at the same time provide more predictable network services. We propose using formal techniques to define and analyze network services, network infrastructure, and interactions between desired services and available infrastructure. The net result will be techniques for predicting the behavior of interacting network services implemented on a heterogeneous, dynamic network infrastructure.
The proposed research activity brings together ongoing research in network design, formal methods, and systems-level design. The fundamental intellectual challenge is the identification of appropriate representations and analysis techniques for network systems within the systems engineering and formal methods domains. The greatest challenge within that task, scaling techniques to the systems level, remains a difficult, unsolved problem. This project will provide sufficient justification for a long- term research program addressing this problem.
Primary Sponsor(s): NSF