In large scale wireless sensor network (WSN) energy reservation is crucial, as in such an environment sensors cannot be periodically maintain. Therefore we investigate the opportunity to reduce the power consumption b...
详细信息
ISBN:
(纸本)9781479950416
In large scale wireless sensor network (WSN) energy reservation is crucial, as in such an environment sensors cannot be periodically maintain. Therefore we investigate the opportunity to reduce the power consumption by reducing the data rate traffic of the network. This is done utilizing either data correlation and sparsity in one dimension or the spatial sparsity among clustered sensor nodes. We found that the data rate can be significantly reduced with minimum recovery error;this extend the life time of the network. Moreover, utilizing the predefined wireless sensing clustering assuming that the nodes in a cluster are sharing most of the sparse supports. Thus, distributed compressive sensing in such a case enhances the whole life time. Finally, we investigated how to adaptively compromise between the measurement error and energy reduction to have a moderate network life time with an accepted error rate.
There is an upsurge in the usage and adaptation of streaming applications in the recent years by both industry and academia. At the core of these applications is streaming data processing engines that perform resource...
详细信息
ISBN:
(纸本)9781538683507
There is an upsurge in the usage and adaptation of streaming applications in the recent years by both industry and academia. At the core of these applications is streaming data processing engines that perform resource management and allocation in order to support continuous track of queries over distributed data streams. Several stream processing engines exists to handle these distributed streaming applications. In this paper, we present different challenges of the stream processing systems, in particular to stateful operators and implement Linear Road benchmark to examine the characteristic and performance metrics of the streaming system, in particular Apache Flink. Furthermore, we examine that Apache Flink can be used as a core for an efficient Linear Road application implementation for distributed environments without breaching the SLA requirements of the application.
For many applications, computation load varies over time. Such applications require dynamic load balancing to improve performance. Centralized load balancing schemes, which perform the load balancing decisions at a ce...
详细信息
ISBN:
(纸本)9781450323789
For many applications, computation load varies over time. Such applications require dynamic load balancing to improve performance. Centralized load balancing schemes, which perform the load balancing decisions at a central location, are not scalable. In contrast, fully distributed strategies are scalable but typically do not produce a balanced work distribution as they tend to consider only local information. This paper describes a fully distributed algorithm for load balancing that uses partial information about the global state of the system to perform load balancing. This algorithm, referred to as GrapevineLB, consists of two stages: global information propagation using a lightweight algorithm inspired by epidemic [21] algorithms, and work unit transfer using a randomized algorithm. We provide analysis of the algorithm along with detailed simulation and performance comparison with other load balancing strategies. We demonstrate the effectiveness of GrapevineLB for adaptive mesh refinement and molecular dynamics on up to 131,072 cores of BlueGene/Q.
This paper proposes a distributed algorithm that aims for congestion control in Software Defined networking with distributed controllers by exploiting loop structures in a network and coloring. Our algorithm dissects ...
详细信息
ISBN:
(纸本)9781479969593
This paper proposes a distributed algorithm that aims for congestion control in Software Defined networking with distributed controllers by exploiting loop structures in a network and coloring. Our algorithm dissects a network into loop units called tie-sets and updates flows on edges by enclosed calculation in each controller, which is in charge of one of loop units. Additionally, the deployment of coloring in a tie-set meta-graph enables the distributed algorithm to avoid update conflicts in shared edges with less communications among controllers.
In this paper, a distributed power control scheme for the third generation wideband CDMA system is presented. In the algorithm, a kind of power update function of exponential type is presented, By adjusting parameter ...
详细信息
ISBN:
(纸本)078039335X
In this paper, a distributed power control scheme for the third generation wideband CDMA system is presented. In the algorithm, a kind of power update function of exponential type is presented, By adjusting parameter k in the power update function, the system can satisfy different performance requirement. Simulation results reveal that this algorithm has faster convergence speed and better system stability than the conventional distributed power control algorithm.
Along with the growth of service type and user number due to network evolution and convergence, higher performance is required for service. logic execution environment (SLEE) in service provisioning. In this paper, an...
详细信息
ISBN:
(纸本)9781424405169
Along with the growth of service type and user number due to network evolution and convergence, higher performance is required for service. logic execution environment (SLEE) in service provisioning. In this paper, an accumulated k-subset algorithm (AK algorithm) is proposed to balance load in distributed SLEE. Based on a model of. resource heterogeneity and load vector, AK algorithm improves k-subset algorithm by accumulating load information within every update interval. Experiments on different update intervals and request arrival rates suggest AK further reduces herd effect due to stale load information, and outperforms k-subset algorithm by 5%-10%.
networking forms an essential part of distributed systems and reliable point-to-point communication is one of the main utilization of distributednetworking systems. In this paper, we discuss new techniques for improv...
详细信息
ISBN:
(纸本)0769520235
networking forms an essential part of distributed systems and reliable point-to-point communication is one of the main utilization of distributednetworking systems. In this paper, we discuss new techniques for improving networking in distributed systems under a sequence of operations, such as inserting and deleting a link between two nodes. As a result, by simply alternating the insertion and deletion of a link, we can generate a sequence of operations that at each step changes the number of components. It is shown how to rest at any time whether two nodes belong to the same sub-network, and how to insert and delete a link in O(m(2/3)) time in the worst case, where m is the current number of links in the network.
In this paper, we describe a new distributed community detection algorithm for billion-edge directed graphs that, unlike modularity-based methods, achieves cluster quality on par with the best-known algorithms in the ...
详细信息
ISBN:
(纸本)9781450337236
In this paper, we describe a new distributed community detection algorithm for billion-edge directed graphs that, unlike modularity-based methods, achieves cluster quality on par with the best-known algorithms in the literature. We show that a simple approximation to the best-known serial algorithm dramatically reduces computation and enables distributed evaluation yet incurs only a very small impact on cluster quality. We present three main results: First, we show that the clustering produced by our scalable approximate algorithm compares favorably with prior results on small synthetic benchmarks and small real-world datasets (70 million edges). Second, we evaluate our algorithm on billion-edge directed graphs (a 1.5B edge social network graph, and a 3.7B edge web crawl), and show that the results exhibit the structural properties predicted by analysis of much smaller graphs from similar sources. Third, we show that our algorithm exhibits over 90% parallel efficiency on massive graphs in weak scaling experiments.
distributed locks are used to guarantee the distributed client-cache coherence in parallel file systems. However, they lead to poor performance in the case of parallel writes under high-contention workloads. We analyz...
详细信息
ISBN:
(纸本)9781665454445
distributed locks are used to guarantee the distributed client-cache coherence in parallel file systems. However, they lead to poor performance in the case of parallel writes under high-contention workloads. We analyze the distributed lock manager and find out that lock conflict resolution is the root cause of the poor performance, which involves frequent lock revocations and slow data flushing from client caches to data servers. We design a distributed lock manager named SeqDLM by exploiting the sequencer mechanism. SeqDLM mitigates the lock conflict resolution overhead using early grant and early revocation while keeping the same semantics as traditional distributed locks. To evaluate SeqDLM, we have implemented a parallel file system called ccPFS using both SeqDLM and traditional distributed locks. Evaluations on 96 nodes show SeqDLM outperforms the traditional distributed locks by up to 10.3x for high-contention parallel writes on a shared file with multiple stripes.
暂无评论