Wireless Computing Architecture II
Abstract
In the next decade we will need large-scale and sophisticated information processing, such as near real-time signal analysis, machine learning and query processing, at the tactical edge that integrate sensor and intelligence collection. This calls for on-demand instantly deployable computing resources to support on-going local operations in the field. Wireless computing and networking technologies that can cope with the relatively modest bandwidths of wireless links while being able to take advantage of their broadcast mature will need to be developed. We have obtained new results capable of addressing aspects of these challenges. First, our methodology of finited-state modeling can provide wireless link models for unmanned aerial vehicles (UAVs) in emulating UAV-based airborne computing systems. Second, our speculative pre-fetching approach hides I/O latency for situations where I/O operations with distant nodes can be a performance bottleneck, as in the case when a UAV interacts with other UAVs or ground nodes.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 2010
- Accession Number
- ADA533077
Entities
People
- H.t. Kung
Organizations
- Harvard University