On the influence of start-up costs in scheduling divisible loads on bus networks

Abstract

AbstractÐOptimal distribution of divisible loads in bus networks is considered in this paper. The problem of minimizing the processing time is investigated by including all the overhead components that could penalize the performance of the system, in addition to the inherent communication and computation delays. These overheads are considered to be constant additive factors to the respective communication and computation components. Closed-form solution for the processing time is derived and the influence of overheads on the optimal processing time is analyzed. We derive a necessary and sufficient condition for the existence of the optimal processing time. We then study the effect of changing the load distribution sequence on the time performance. Through rigorous analysis, an optimal sequence to distribute the load among the processors is identified, whenever it exists. In case such an optimal sequence fails to exist, we present a greedy algorithm to obtain a suboptimal sequence based on some important properties of the overhead factors. Then, the effect of granularity of the data that is divisible is considered in the analysis for the case of homogeneous networks. An integer approximation algorithm capable of generating integer values of the load fractions in time ym, where m is the number of processors in the network, is proposed. We then show that the upper bound on the suboptimal solution generated by our algorithm lies within a radius given by the sum of the computation and communication delays. Several numerical examples are presented to illustrate the concepts.

Algorithms for distributed feature extraction in multi-camera visual sensor networks

“…Usually three decisions need to be made: the subset of the processors used, the order they receive their share of workload, and the division of the workload. Unfortunately, the results are specific to a given system setup, and therefore scheduling solutions are derived for given, simplified systems with simple topologies [29], [30]. In [23] we show that the application of DLT for distributed visual processing is non-trivial even in the case of a single sensor node, due to the transmission overhead introduced by distributed feature extraction, and due to the dynamism of the frame content in the video.…”

Cooperative image analysis in visual sensor networks

“…While the optimal scheduling order is in the decreasing order of link transmission rates in many basic scenarios, as derived in [20], in the presence of transmission overlap the optimal order depends also on the processing rates of the nodes [32], and no general results can be derived for networks with heterogeneous transmission and processing rates. Therefore, in the numerical evaluation we consider homogeneous processing rates, where the basic result on the optimal scheduling order holds.…”

Divisible Load Distribution in a Network of Processors

“…The concept of equivalent processor is first introduced in [25][26] and lately exploited by [27]. The idea is to replace a set of k processors by a single processor, which has equivalent computing power and an equivalent distance from the root processor.…”