A Comprehensive Worst-Case Calculus for Wireless Sensor Networks with In-Network Processing

A calculus for information-driven networks

“…Particularly, the spatial and temporal correlations of information have been studied and utilized in network scheduling and resource saving in wireless sensor networks [16], [18], [20]; information redundancy has been exploited to help load balancing and improve fault tolerance in peer-to-peer content sharing systems [2], [12]. Regarding the third challenge, to the best of our knowledge, the only attempts to accommodate information processing in performance modeling are the work in [8], [19]. Nevertheless, information processing is simply modeled with a scaling function in [8], [19] and the information embedded in data packets has not been modeled, let alone utilized.…”

“…Regarding the third challenge, to the best of our knowledge, the only attempts to accommodate information processing in performance modeling are the work in [8], [19]. Nevertheless, information processing is simply modeled with a scaling function in [8], [19] and the information embedded in data packets has not been modeled, let alone utilized. In this sense, the performance models in [8], [19] are not really information-driven models.…”

“…Nevertheless, information processing is simply modeled with a scaling function in [8], [19] and the information embedded in data packets has not been modeled, let alone utilized. In this sense, the performance models in [8], [19] are not really information-driven models. A systematic performance modeling framework suitable for information-driven networks still remains largely open.…”

“…Nevertheless, they are different according to Remark 3. In information-driven networks, in-network information processing is likely performed. From this in-network processing viewpoint, the present paper is related to [8] and [19]. In [8], scaling functions are used to model the relationship between the input traffic and the output traffic of a network element that processes the traffic.…”

“…Based on the proposed scaling server model, [8] extends the deterministic network calculus by considering data scaling in networks with innetwork data processing. In [19], how the scaling elements can be shifted across multiplexers in sensor networks is studied, which enables worst-case analysis of traffic delay and backlog in such networks. Note that in [8], [19], information processing only applies to intra-flow data, leaving inter-flow processing un-considered.…”