Processing streams of data in an overlay network of operators distributed over a wide-area network is a common idea shared by different applications such as distributed event correlation systems and large-scale sensor...
详细信息
Processing streams of data in an overlay network of operators distributed over a wide-area network is a common idea shared by different applications such as distributed event correlation systems and large-scale sensor networks. In order to utilize network resources efficiently and allow for the parallel deployment of a large number of large-scale operator networks, suitable placement algorithms are vital that place operators on physical nodes. In this paper, we present a distributed placement algorithm that minimizes the bandwidth-delay product of data streams between operators of the network in order to reduce the induced network load. Since the fundamental optimization problem is NP-hard, we propose a heuristic solution. First, we calculate an optimal solution in an intermediate continuous search space, called latency space. Subsequently the continuous solution is mapped to the physical network. Our evaluations show that this algorithm reduces the resulting network load significantly compared to state of the art algorithms and achieves results close to the optimum.
In the context of the prosperous development of Proteomics in life science, protein quantification, especially these based on Mass Spectrometry (short for MS) method, becomes an essential part of research. In our prev...
详细信息
the proceedings contain 57 papers. the topics discussed include: logical topology design with traffic grooming for scheduled demands;the impact of multi-core architectures on task retrieval policies for volunteer comp...
ISBN:
(纸本)9780889867741
the proceedings contain 57 papers. the topics discussed include: logical topology design with traffic grooming for scheduled demands;the impact of multi-core architectures on task retrieval policies for volunteer computing;evaluation of CPU power control and scheduling technique on parallelcomputing environment;clustered assignment of CCQS in information monitoring system supporting conjunctive queries;dense skip graphs as an overlay for P2P networks;multi-connection and multi-core aware all-gather on infiniband clusters;an improved prioritization method for task scheduling in heterogeneous computingsystems;thread pipelining with clustered data communication on commodity CMPS;a distributed algorithm for disjoint paths in star networks;parallel algorithms for a visual text mining platform;and a cache miss analysis for multithreaded architectures.
the results of application of an effective method of a search for global extremum for a problem solution of optimization of parameters of waveguide structures by specified electrodynamics characteristics on systems of...
详细信息
Large-scale, parallel clusters composed of commodity processors are increasingly available, enabling the use of vast processing capabilities and distributed RAM to solve hard search problems. We investigate parallel a...
详细信息
the results of application of an effective method of a search for global extremum for a problem solution of optimization of parameters of waveguide structures by specified electrodynamics characteristics on systems of...
详细信息
the results of application of an effective method of a search for global extremum for a problem solution of optimization of parameters of waveguide structures by specified electrodynamics characteristics on systems of distributed and parallel calculations are presented.
暂无评论