Operating Environment on-line metrics for application architecture matching

The need to execute highly demanding and real-time applications on ambient computing networks will be a big challenge in the future. In this paper we present an operating environment (OE) that allows application partitioning on such platforms. The OE decisions are made based on a trade negotiation protocol and energy efficiency and resource matching metrics. The metrics matches application and architecture using purely application and architectural defined characteristics. Application is modeled as hierarchical task graphs with meta-data and architecture as a reconfigurable and multi-granular instruction set virtual processor. Metrics complexity is adequate for on-line decisions thanks to easy generalization through the application hierarchy.