Results 1  10
of
15
Impact of Failure on Interconnection Networks for Large Storage Systems
 IN PROCEEDINGS OF THE 22ND IEEE / 13TH NASA GODDARD CONFERENCE ON MASS STORAGE SYSTEMS AND TECHNOLOGIES
, 2005
"... Recent advances in largecapacity, lowcost storage devices have led to active research in design of largescale storage systems built from commodity devices for supercomputing applications. Such storage systems, composed of thousands of storage devices, are required to provide high system bandwidth ..."
Abstract

Cited by 5 (1 self)
 Add to MetaCart
Recent advances in largecapacity, lowcost storage devices have led to active research in design of largescale storage systems built from commodity devices for supercomputing applications. Such storage systems, composed of thousands of storage devices, are required to provide high system bandwidth and petabytescale data storage. A robust network interconnection is essential to achieve high bandwidth, low latency, and reliabile delivery during data transfers. However, failures, such as temporary link outages and node crashes, are inevitable. We discuss the impact of potential failures on network interconnections in very largescale storage systems and analyze the tradeoffs among several storage network topologies by simulations. Our results suggest that a good interconnect topology be essential to faulttolerance of a petabytescale storage system.
Explicit Constructions of Fault Tolerant Open Linear Arrays
, 2005
"... Two graph models for a Kfault tolerant linear array with external inputs and outputs are proposed: one is with minimum number of spare nodes and incurs low spare resource overhead, and the other is with small internal degree and incurs small runtime reconfiguration hardware overhead. Both graph mo ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Two graph models for a Kfault tolerant linear array with external inputs and outputs are proposed: one is with minimum number of spare nodes and incurs low spare resource overhead, and the other is with small internal degree and incurs small runtime reconfiguration hardware overhead. Both graph models can be applied to fault tolerant VLSI designs while maintaining low hardware cost.
Authenticated communication and computation in knowntopology networks with a trusted authority
, 2009
"... not be interpreted as representing the official policies, either expressed or implied, of any sponsoring institution, the U.S. government or any other entity. Keywords: secure aggregation, sensor networks, data aggregation, authenticated communication, known topology networksPhD. Dissertation: Haowe ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
not be interpreted as representing the official policies, either expressed or implied, of any sponsoring institution, the U.S. government or any other entity. Keywords: secure aggregation, sensor networks, data aggregation, authenticated communication, known topology networksPhD. Dissertation: Haowen Chan We show that two distinguishing properties of sensor networks, i.e., the presence of a trusted base station, and the preknowledge of the fixed network topology, can yield security protocols that are both communicationefficient and highly general. We show new protocols for broadcast authentication, credential dissemination and nodetonode signatures. For securing innetwork distributed computations, we show an algorithm for securely computing the sum of sensor readings in the network, which we can generalize to tree computations for any combination of continuous realvalued functions. Each of these primitives involves pernode communication costs that scale logarithmically with the number of nodes in the network, do not require public key cryptography, and are secure against arbitrary coalitions of malicious nodes. The broadcast authentication scheme achieves better properties with fewer assumptions than existing work, and the other new protocols
SelfHealing Asynchronous Arrays
"... This paper presents a systematic method for designing of a selfhealing asynchronous array in the presence of errors. By adding spare resources in one of three different ways and forcing the asynchronous circuit to stall in case of failure, the specific selfreconfiguration logic is activated by a d ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
This paper presents a systematic method for designing of a selfhealing asynchronous array in the presence of errors. By adding spare resources in one of three different ways and forcing the asynchronous circuit to stall in case of failure, the specific selfreconfiguration logic is activated by a deadlock detector and the array circuit can be reconfigured around the faulty components and recover from errors automatically. Experimental evaluations show that this method requires less hardware cost, smaller critical circuit size, lower performance overhead and is more scalable than traditional NMRbased techniques.
SiameseTwin: A Dynamically Faulttolerant Fattree
"... Fattrees are a special case of multistage interconnection networks with quite good static fault tolerance capabilities. They are however straightforwardly unable to provide local dynamic fault tolerance. In this paper we propose a network topology based on the fattree using two parallel networks w ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Fattrees are a special case of multistage interconnection networks with quite good static fault tolerance capabilities. They are however straightforwardly unable to provide local dynamic fault tolerance. In this paper we propose a network topology based on the fattree using two parallel networks with crossover links between them in an effort to enable dynamic fault tolerance. We evaluate and compare this topology with two other similar fattree topologies and show through simulations that the new topology is able to improve slightly upon the ability to tolerate faults statically. More importantly, we show that the new network topology is the only one of the evaluated topologies able to tolerate one fault dynamically, with a superior network performance in the face of dynamically handled faults. 1.
Hierarchical Topological Network Design
"... Abstract—We present a hierarchical solution method to approximately solve the topological network design problem: given positive integers ( 1), minimize the number of arcs required to interconnect nodes, so that the network diameter does not exceed, the maximum node degree does not exceed 1, and the ..."
Abstract
 Add to MetaCart
Abstract—We present a hierarchical solution method to approximately solve the topological network design problem: given positive integers ( 1), minimize the number of arcs required to interconnect nodes, so that the network diameter does not exceed, the maximum node degree does not exceed 1, and the network is single node survivable. The method uses dynamic programming to piece together small networks to create larger networks. The method was used to plan two highspeed packet networks at AT&T. Index Terms—Dynamic programming, hierarchical design, network design, survivability. I. INTRODUCTION AND PROBLEM DEFINITION PLANNING the evolution of a large communications network requires a variety of approximations and estimates. A demand forecast for the various services is required, as well as lowlevel service planning assumptions (e.g., card redundancy).
SANTA CRUZ UNDERSTANDING AND COPING WITH FAILURES IN LARGESCALE STORAGE SYSTEMS
, 2005
"... by ..."
Categories and Subject Descriptors C.1.4 [Computer Systems Organization]: Processor Architectures—parallel
"... This paper proves that for every positive integers n and k, we can explicitly construct a graph G with n+O(k) vertices and maximum degree 3, such that even after removing any k vertices from G, the remaining graph still contains a path of length n − 1. This settles a problem raised by Zhang [11,12] ..."
Abstract
 Add to MetaCart
This paper proves that for every positive integers n and k, we can explicitly construct a graph G with n+O(k) vertices and maximum degree 3, such that even after removing any k vertices from G, the remaining graph still contains a path of length n − 1. This settles a problem raised by Zhang [11,12] in connection with the design of faulttolerant linear arrays.
Fault Tolerant Asynchronous Adder through Dynamic Selfreconfiguration
"... This paper presents a systematic method for the design of a selfhealing asynchronous adder. We propose a graphbased model for the design of a faulttolerant linear array with external inputs and outputs with a minimum number of spare resources. A Kfaulttolerant asynchronous adder design is presen ..."
Abstract
 Add to MetaCart
This paper presents a systematic method for the design of a selfhealing asynchronous adder. We propose a graphbased model for the design of a faulttolerant linear array with external inputs and outputs with a minimum number of spare resources. A Kfaulttolerant asynchronous adder design is presented based on this analysis, together with the necessary support logic for dynamic selfreconfiguration. Experimental evaluations show that our method incurs both low hardware cost and small performance overhead compared to traditional approaches to faulttolerance. 1
unknown title
"... This paper presents a systematic method for designing of a selfhealing asynchronous array in the presence of errors. By adding spare resources in one of three different ways and forcing the asynchronous circuit to stall in case of failure, the specific selfreconfiguration logic is activated by a d ..."
Abstract
 Add to MetaCart
This paper presents a systematic method for designing of a selfhealing asynchronous array in the presence of errors. By adding spare resources in one of three different ways and forcing the asynchronous circuit to stall in case of failure, the specific selfreconfiguration logic is activated by a deadlock detector and the array circuit can be reconfigured around the faulty components and recover from errors automatically. Experimental evaluations show that this method requires less hardware cost, smaller critical circuit size, lower performance overhead and is more scalable than traditional NMRbased techniques. 1