KSEF R&D Excellence: Enhancing Network Layer Services

The Internet's network layer is critical to interoperability, and thus the barriers to changing it are understandably high. As a result it is notoriously difficult to deploy new, advanced services in the network infrastructure. Service provides will not even consider changes unless there is a strong business case to be made. Unfortunately this results in a "chicken-and-egg" problem: service providers are nOt likely to invest resources to deploy a new service until there exist lucrative applications that depend on the service. On the other hand, these new applications cannot be written, tested, or demonstrated if the service does not exist. The problem is even worse for services that must be supported end-to-end (i.e, services that must be deployed everywhere) before applications can benefit. Well-known (failed) examples of such services indude multicast and guaranteed or "premium" performance services. We have developed two router-based building-block services, ephemeral state processing (ESP) and light;veight processing modules (LWP), that offer extended network-layer functionality, can be deployed non-intrusively as add-ons to the existing network layer, offer benefits even when partially deployed, and can operate at current and emerging network line speeds. The goal of this project is to demonstrate that these building blocks can be efficiently implemented, deployed, and used to enable new advanced applications and services. Our initial results are very encouraging. We will refine and extend our building block services, and investigate their scalability and performance limits through experimentation and implementation. We will also develop new applications to demonstrate the advantages of our approach. To gain deployment experience and demonstrate the ability to operate in a production network, we will deploy and experiment with the service flrst locally, and then in a production network.