RAPLang

Abstract

In active networks, active packet can retrieve information of intermediate nodes for packet processing and services are dynamically deployed. These features are suitable for network configuration that causes heavy traffic and uses information of network environment. In this study, we propose a Resource-Aware Packet Language (RAPLang) to replace heavy-traffic and time-consuming network configuration with active packet-based and automated network configuration in active networks. RAPLang provides resource-aware instructions that decide whether to process packet based on the amount of available resources at active node. It is useful for gathering and analyzing information that is required for network configuration and applying it to packet processing. Furthermore, we optimize interpreter for these instructions. The performance evaluation results show that the proposed approach reduces average configuration time and average interpreting time for network configuration as much as 30% compared with the PLAN-based approach.