Search

Broadcast is a widely used operation in many streaming and deep learning applications to disseminate large amounts of data on emerging heterogeneous High-Performance Computing (HPC) systems. However, traditional broadcast schemes do not fully utilize hardware features for Graphics Processing Unit (GPU)-based applications. In this paper, a model-oriented analysis is presented to identify performance bottlenecks of existing broadcast schemes on GPU clusters. Next, streaming-based broadcast schemes are proposed to exploit InfiniBand hardware multicast (IB-MCAST) and NVIDIA GPUDirect technology for efficient message transmission. The proposed designs are evaluated in the context of using Message Passing Interface (MPI) based benchmarks and applications. The experimental results indicate improved scalability and up to 82 percent reduction of latency compared to the state-of-the-art solutions in the benchmark-level evaluation. Furthermore, compared to the state-of-the-art, the proposed design yields stable higher throughput for a synthetic streaming workload, and 1.3x faster training time for a deep learning framework. [ABSTRACT FROM AUTHOR]

XML data broadcast is an efficient way to disseminate semistructured information in wireless mobile environments. In this paper, we propose a novel two-tier index structure to facilitate the access of XML document in an on-demand broadcast system. It provides the clients with an overall image of all the XML documents available at the server side and hence enables the clients to locate complete result sets accordingly. A pruning strategy is developed to cut down the index size and a two-tier structure is proposed to further remove any redundant information. In addition, two index distribution strategies, namely naive distribution and partial distribution, have been designed to interleave the index information with the XML documents in the wireless channels. Theoretical analysis and simulation experiments are also put forward to show the benefits of our indexing methods. [ABSTRACT FROM AUTHOR]

In this paper, we consider the issue of data broadcasting in mobile social networks (MSNets). The objective is to broadcast data from a superuser to other users in the network. There are two main challenges under this paradigm, namely 1) how to represent and characterize user mobility in realistic MSNets; 2) given the knowledge of regular users' movements, how to design an efficient superuser route to broadcast data actively. We first explore several realistic data sets to reveal both geographic and social regularities of human mobility, and further propose the concepts of geocommunity and geocentrality into MSNet analysis. Then, we employ a semi-Markov process to model user mobility based on the geocommunity structure of the network. Correspondingly, the geocentrality indicating the “dynamic user density” of each geocommunity can be derived from the semi-Markov model. Finally, considering the geocentrality information, we provide different route algorithms to cater to the superuser that wants to either minimize total duration or maximize dissemination ratio. To the best of our knowledge, this work is the first to study data broadcasting in a realistic MSNet setting. Extensive trace-driven simulations show that our approach consistently outperforms other existing superuser route design algorithms in terms of dissemination ratio and energy efficiency. [ABSTRACT FROM PUBLISHER]

In an unstructured peer-to-peer (P2P) network (e.g., Gnutella), participating peers choose their neighbors randomly such that the resultant P2P network mismatches its underlying physical network, resulting in the lengthy communication between the peers and redundant network traffics generated in the underlying network. Previous solutions to the topology mismatch problem in the literature either have no performance guarantees or are far from the optimum. In this paper, we propose a novel topology matching algorithm based on the Metropolis-Hastings method. Our proposal is guided by our insight analytical model and is close to the optimal design. Specifically, we show that our proposal constructs an unstructured P2P network in which a broadcast message, originated by any node v, reaches any other node u by taking approximately the only physical end-to-end delay between v and u. In addition, our design guarantees the exponential broadcast scope. We verify our solution through extensive simulations and show that our proposal considerably outperforms state-of-the-art solutions. [ABSTRACT FROM AUTHOR]

Two-layer hierarchy unstructured peer-to-peer (P2P) systems, comprising an upper layer of superpeers and an underlying layer of ordinary peers, are commonly used to improve the performance of large-scale P2P systems. However, the optimal superpeer network design involves several requirements including superpeer degree, network diameter, scalability, load balancing, and flooding performance. A perfect difference graph has desirable properties to satisfy the above design rationale of superpeers overlay network. This paper proposes a two-layer hierarchical unstructured P2P system in which a perfect difference graph (PDG) is used to dynamically construct and maintain the superpeer overlay topology. In addition, the broadcasting performance of the P2P system is enhanced through the use of a PDG-based forwarding algorithm, which ensures that each superpeer receives just one lookup query flooding message. The theoretical results show that the proposed system improves existing superpeer hierarchical unstructured P2P systems in terms of a smaller network diameter, fewer lookup flooding messages, and a reduced average delay, and the experimental results show that the proposed two-layer hierarchy P2P system performs very well in the dynamic network environment. [ABSTRACT FROM AUTHOR]

In this paper, we consider the issue of efficient broadcasting in mobile ad hoc networks (MANET5) using network coding and directional antennas. Network coding-based broadcasting focuses on reducing the number of transmissions each forwarding node performs in the multiple source/multiple message broadcast application, where each forwarding node combines some of the received messages for transmission. With the help of network coding, the total number of transmissions can be reduced compared to broadcasting using the same forwarding nodes without coding. We exploit the usage of directional antennas to network coding-based broadcasting to further reduce energy consumption. A node equipped with directional antennas can divide the omnidirectional transmission range into several sectors and turn some of them on for transmission. In the proposed scheme using a directional antenna, forwarding nodes selected locally only need to transmit broadcast messages, original or coded, to restricted sectors. We also study two extensions. The first extension applies network coding to both dynamic and static forwarding node selection approaches. In the second extension, we design two approaches for the single source/single message issue in the network coding-based broadcast application. Performance analysis via simulations on the proposed algorithms using a custom simulator and ns2 is presented. [ABSTRACT FROM AUTHOR]

Prior studies show that more than 70 percent of communication paths in a popular unstructured peer-to-peer (P2P) system (i.e., Gnutella) do not exploit the physical network topology, leading to the topology mismatch problem, and thus, lengthen communication between participating peers. While previous efforts in solving overlay topology matching problems do not guarantee the bounds of performance metrics (e.g., the communication delay between any two overlay peers and the broadcasting scope of any participating peer), in this paper, we present a novel topology matching algorithm that has provable performance qualities. In our proposal, each participating node creates and manages a constant number of overlay connections to other peers in a distributed manner. In rigorous performance analysis, we show that 1) the expected overlay communication delay between any two nodes in our P2P network is a constant; 2) in addition, any joining node has the exponential broadcasting scope in expectation; 3) furthermore, a participating node takes a polylogarithmic overhead to exploit the physical network locality and maintain its flooding scope. Together with extensive simulations, we present our proposal that significantly outperforms two recent solutions, i.e., THANCS and mOverlay, in terms of overlay communication latency and/or broadcasting scope. [ABSTRACT FROM AUTHOR]

In this paper, we consider the communications involved by the execution of a complex application, deployed on a heterogeneous platform. Such applications extensively use macrocommunication schemes, for example, to broadcast data items. Rather than aiming at minimizing the execution time of a single broadcast, we focus on the steady-state operation. We assume that there is a large number of messages to be broadcast in pipeline fashion, and we aim at maximizing the throughput, i.e., the (rational) number of messages which can be broadcast every time-step. We target heterogeneous platforms, modeled by a graph where resources have different communication and computation speeds. Achieving the best throughput may well require that the target platform is used in totality: We show that neither spanning trees nor DAGs are as powerful as, general graphs. We show how, to compute the best throughput using linear programming, and how to exhibit a periodic schedule, first when restricting to a DAG, and then when using a general graph. The polynomial compactness of the description comes from the decomposition of the schedule into several broadcast trees that are used concurrently to reach the best throughput. It is important to point out that a concrete scheduling algorithm based upon the steady-state operation is asymptotically optimal, in the class of all possible schedules (not only periodic solutions). [ABSTRACT FROM AUTHOR]

Due to advances in fiber-optics and VLSI technology, interconnection networks that allow multiple simultaneous broadcasts are becoming feasible. Distributed-shared-memory implementations on such networks promise high performance even for applications with small granularity. This paper presents the architecture of one such implementation, called the Simultaneous Optical Multiprocessor Exchange Bus, and examines the performance of augmented DSM protocols that exploit the natural duplication of data to maintain a recovery memory in each processing node and provide basic fault tolerance. Simulation results show that the additional data duplication necessary to create fault-tolerant DSM causes no reduction in system performance during normal operation and eliminates most of the overhead at checkpoint creation. Under certain conditions, data blocks that are duplicated to maintain the recovery memory are utilized by the underlying DSM protocol, reducing network traffic, and increasing the processor utilization significantly. [ABSTRACT FROM AUTHOR]

In this paper, we propose a family of structures, namely, κ-localized minimum spanning tree (LMSTκ) for topology control and broadcasting in wireless ad hoc networks. We give an efficient localized method to construct LMSTκ using only O(n) messages under the local-broadcast communication model, i.e., the signal sent by each node will be received by all nodes within the node's transmission range. We also analytically prove that the node degree of the structure LMSTκ is at most 6, LMSTκ is connected and planar and, more importantly, the total edge length of the LMSTk is within a constant factor of that of the minimum spanning tree when κ ≥ 2 (called low weighted hereafter). We then propose another low weighted structure, called Incident MST and RNG Graph (1MRG), that can be locally constructed using at most 13n messages under the local broadcast communication model. Test results are corroborated in the simulation study. We study the performance of our structures in terms of the total power consumption for broadcasting, the maximum node power needed to maintain the network connectivity. We theoretically prove that our structures are asymptotically the best possible for broadcasting among all locally constructed structures. Our simulations show that our new structures outperform previous locally constructed structures in terms of broadcasting and power assignment for connectivity. [ABSTRACT FROM AUTHOR]

Self-pruning is an effective scheme for efficient broadcasting in ad hoc wireless networks. In a self-pruning broadcast protocol, a node may not forward a broadcast packet if a certain self-pruning condition is satisfied based on the neighborhood information. In a static network with an ideal MAC layer, only a subset of nodes forward the broadcast packet and still guarantee the complete network delivery. Various protocols have been proposed with different self-pruning conditions. Recently, a generic self-pruning protocol was proposed by Wu and Dai [21], which combines the strength of previous conditions and is more effective. In this paper, we first propose an enhanced version of the generic protocol, which is more elegant in interpreting existing protocols and has a simpler correctness proof. Then, we evaluate the performance of the family of self-pruning protocols under various network situations with ns2. The objective is to observe the efficiency and reliability of these protocols as a function of network density, congestion, and mobility, and provide a guideline of implementation in the "real world." Our performance analysis reveals that the protocol reliability is barely affected by packet collision. However, most self-pruning protocols suffer from low delivery ratio in highly mobile networks. We further explore various techniques that improve the delivery ratio and show that both high efficiency and reliability can be achieved in highly mobile networks. [ABSTRACT FROM AUTHOR]

This paper presents a scalable and adaptive decentralized metadata lookup scheme for ultralarge-scale file systems (more than Petabytes or even Exabytes). Our scheme logically organizes metadata servers (MDSs) into a multilayered query hierarchy and exploits grouped Bloom filters to efficiently route metadata requests to desired MDSs through the hierarchy. This metadata lookup scheme can be executed at the network or memory speed, without being bounded by the performance of slow disks. An effective workload balance method is also developed in this paper for server reconfigurations. This scheme is evaluated through extensive trace-driven simulations and a prototype implementation in Linux. Experimental results show that this scheme can significantly improve metadata management scalability and query efficiency in ultralarge-scale storage systems. [ABSTRACT FROM AUTHOR]

Our paper [1] generated a lot of interest among researchers in ad hoc networks. A number of researchers questioned, through their articles, or directly to the first author, the correctness of the described procedure, and the correctness of the claim that the procedure does not need any communication exchange between nodes, in addition to "hello" messages needed to learn information about neighboring nodes. This correspondence completes the article by providing the actual dominating set definitions used in the procedure (from which zero communication overhead follows easily), the correct procedure (the published one has few misprints at key places), and the proof that the new definitions and procedure indeed define connected dominating sets. [ABSTRACT FROM AUTHOR]