MapReduce is the renowned distributed and parallel programming model with an extensive support for large scale computing. However, the performance of MapReduce is currently limited by the default scheduler as it is no...
详细信息
ISBN:
(纸本)9781509030385
MapReduce is the renowned distributed and parallel programming model with an extensive support for large scale computing. However, the performance of MapReduce is currently limited by the default scheduler as it is not suitable for heterogeneous environments and it does not consider certain user constraints such as deadlines. The paper proposes a deadline aware scheduling algorithm that selectively uses speculative execution when the job approaches its deadline in order to expedite job's execution. The algorithm is implemented on the heterogeneous Hadoop cluster and the evaluation shows significant improvement in the performance. The performance improvement was observed as the number of jobs that miss the deadlines as well as overall execution time for different workloads was minimized
This paper designs a distributed algorithm for solving sparse semidefinite programming (SDP) problems, based on the alternating direction method of multipliers (ADMM). It is known that exploiting the sparsity of a lar...
详细信息
ISBN:
(纸本)9781479978878
This paper designs a distributed algorithm for solving sparse semidefinite programming (SDP) problems, based on the alternating direction method of multipliers (ADMM). It is known that exploiting the sparsity of a large-scale SDP problem leads to a decomposed formulation with a lower computational cost. The algorithm proposed in this work solves the decomposed formulation of the SDP problem using an ADMM scheme whose iterations consist of two subproblems. Both subproblems are highly parallelizable and enjoy closed-form solutions, which make the iterations computationally very cheap. The developed numerical algorithm is also applied to the SDP relaxation of the optimal power flow (OPF) problem, and tested on the IEEE benchmark systems.
A distributed algorithm is proposed for detection and resolution of resource deadlocks in object-oriented distributed systems. The algorithm can be used in conjunction with concurrency control algorithms that are base...
详细信息
A distributed algorithm is proposed for detection and resolution of resource deadlocks in object-oriented distributed systems. The algorithm can be used in conjunction with concurrency control algorithms that are based on the semantic lock model. To drastically reduce message traffic, the algorithm properly identifies and eliminates redundant messages. It is shown that its worst and average time complexities are O(ne), where e is the number of edges in the waits-for graph and n is the number of vertices.< >
A distributed system is a collection of computers. These computers can work together, but all systems are independent. A leader's election is a basic necessity for distributed systems. When a system is chosen as a...
详细信息
ISBN:
(纸本)9781467352321
A distributed system is a collection of computers. These computers can work together, but all systems are independent. A leader's election is a basic necessity for distributed systems. When a system is chosen as a leader, it should operate as a system's management;make final decisions and the like. In this paper we describe a distributed system and models of it, explain the famous leader election algorithm, or software agent, and we illustrate our algorithm to find a leader in a distributed WSN (Wireless Sensor Network) using software agents, also after these, we will measure time for find leader by use of different number of agents. The simulation result shows that the proposed algorithm can find faster the leader in WSN.
Heterogeneous deployment of base stations with different transmit power levels helps increase the network efficiency. Yet, the allocation of powers to base stations should be optimal in order to minimize interferences...
详细信息
Heterogeneous deployment of base stations with different transmit power levels helps increase the network efficiency. Yet, the allocation of powers to base stations should be optimal in order to minimize interferences and thus improve users through-put. In this paper, we present a game theoretical approach to maximize the overall throughput in heterogeneous networks with optimal power allocation. We formulate the allocation problem as a "potential game". This type of games is characterized by a potential function, and has a specific property: a pure Nash equilibrium solution always exists and it is, moreover, either a local or a global optimizer of the potential function. We use a fully distributed algorithm that requires minimal coordination between base stations to reach a Nash equilibrium, and we prove that it is necessarily a global maximum. The performance of the algorithm is evaluated through simulations.
Constructing a connected dominating set (CDS) has been widely used as a topology control method and a virtual backbone to reduce energy consumption and communication overhead. Previous works emphasize on constructing ...
详细信息
Constructing a connected dominating set (CDS) has been widely used as a topology control method and a virtual backbone to reduce energy consumption and communication overhead. Previous works emphasize on constructing the minimum CDS (MCDS) and do not consider node mobility and CDS reconstruction where energy balancing is required. In this paper, we propose a CDS reconstruction algorithm which is able to reconfigure the CDS locally in the presence of mobility and energy balancing without the need for reconstructing the whole CDS. The algorithm does not increase the approximation factor of the algorithm which has constructed the CDS already and it has O(Delta log Delta) and O(Delta) time and message complexity, respectively, which is a great improvement in comparison with the other approaches, where Delta is the maximum degree of the network graph.
This paper studies correctness and robustness properties of motion coordination algorithms with respect to link failures in the network topology. The technical approach relies on computational geometric tools such as ...
详细信息
This paper studies correctness and robustness properties of motion coordination algorithms with respect to link failures in the network topology. The technical approach relies on computational geometric tools such as proximity graphs, nondeterministic systems defined via set-valued maps and Lyapunov stability analysis. The manuscript provides two general results to help characterize the asymptotic behavior of spatially distributed coordination algorithms. These results are illustrated in rendezvous and flocking coordination algorithms
Large classes of autonomic (self-managing, self-healing) systems can be created by logically integrating simpler autonomic systems. The configuration method is widely used for such integration. However, there are few ...
详细信息
Large classes of autonomic (self-managing, self-healing) systems can be created by logically integrating simpler autonomic systems. The configuration method is widely used for such integration. However, there are few formalized tools in support of this method for specification, compilation, diagnosis, reasoning, and distributed provisioning. As a result, the practice of this method is very costly and can lead to security failures. This paper presents a technique called Service Grammar for building these tools based on a novel analysis of protocols and distributed algorithms in a domain of interest. The technique is illustrated in the context of a realistic adaptive virtual private network. We show how lower-layer adaptive protocols can be composed to create adaptive behavior at a higher layer.
In this paper we present a distributed algorithm to study the benefits of using sparse grooming in IP-over-WDM networks. The algorithm implements a grooming approach that allows the efficient use of resources while al...
详细信息
In this paper we present a distributed algorithm to study the benefits of using sparse grooming in IP-over-WDM networks. The algorithm implements a grooming approach that allows the efficient use of resources while alleviating the management of large and complex topologies.
A set S is a dominating set (DS) if each node in the graph is either in S or a neighbor to at least one of the nodes in S. A connected dominating set (CDS) is a DS that induces a connected subgraph. A t-spanner of a g...
详细信息
A set S is a dominating set (DS) if each node in the graph is either in S or a neighbor to at least one of the nodes in S. A connected dominating set (CDS) is a DS that induces a connected subgraph. A t-spanner of a graph G = (V,E) is a spanning subgraph G' = (V,E'), such that the shortest hop path between any two nodes in G', is at most t times their shortest path in G. A sparse spanner (spanner with linear edges) is of fundamental importance to distributed networking operations. In this paper, we present a new algorithm for constructing and maintaining a CDS-based sparse spanner for mobile ad hoc networks without using geographic positions. This CDS has a constant approximation factor. Consequently, the number of nodes responsible for routing is also within a constant factor of the minimum. Our distributed algorithm runs in linear time and uses linear messages. Furthermore, the spanner has a constant topological and geometric dilation.
暂无评论