An Examination of Routing Algorithms for Parallel Computing Environments

Size: px
Start display at page:

Download "An Examination of Routing Algorithms for Parallel Computing Environments"

Transcription

1 A case can be made that the Achilles heel of parallel processing networks and clusters is that they all have to deal with the unavoidable problem of communication over the System Area Network. In distributed computing networks computers (nodes) are connected in a System Area Network (SANs). Communications between nodes are inherently slower than communications inside a node due to factors as simple as distance, power, and communications medium, and thus can slow down an entire calculation if one node is dependent on receiving data from another one. Complex calculations may be performed faster if you have more than one node working on the problem at the same time, but if the necessary communications between those nodes take too long, the most powerful processors in the world, running parallel in the most powerful computers in the world, running in parallel clusters are not going to be giving you your results any quicker. Thus, a great deal of effort in high-performance parallel-network architecture is geared towards optimizing these inter-node communications, to increase their speed and reduce the overhead that can bring down a potentially faster program. Since not much can be done about power and communications medium, we programmers tend to focus our optimizing efforts on path-selection algorithms, or algorithms to determine the most ideal route to get data (i.e. a message) from one node in a network to another. Optimizing the path selection algorithm for lower latency and increased throughput will decrease the overall execution time required to run programs on a distributed computing network. The shorter and quicker (two different concepts which will be explained below) the path a given message takes, the less time is spent by the receiving node waiting around, and thus slowing up the final result. Shorter paths

2 invariably are as close to a straight-line path as possible from source to destination, while quicker paths take the least time. As any person who has had to sit in half and hour of traffic to travel three blocks knows shorter is not always quicker, and vice versa. And this is where a lot of path-selection algorithms focus most of their efforts, in avoiding those traffic jams which threaten to derail the quickness of a shortest-path route, while not sending a message on an unnecessarily circuitous route to its destination because it has less of a chance of being clogged. These traffic jams can be caused by congestion (other nodes, sending other messages over the same routes) or deadlocks (where two or more messages are basically waiting infinitely for the other one to pass). Deterministic algorithms plot an entire route in advance, typically using global network-state data, in order to avoid those congestions. Adaptive algorithms typically need little to no global network-state data, and instead change their routes on-the-fly depending on conditions they encounter. Some algorithms can take advantage of Virtual Channels (VC), if the hardware and algorithm allow, catching deadlocks ahead of time. We are going to be taking a look at a few path-selection algorithms designed for routing data and messages in networks, especially parallel-processing clusters. These algorithms run the spectrum in the different ways to approach the path-selection problem, tackling regular and irregular topologies, deterministic and adaptive routing, even virtual channels or the lack thereof. One huge hurdle in plotting out ideal paths for messages is the topography of the network itself. Regular networks correspond to some sort of geographical graph structure (1), and thus are easier to deterministically plot out. Typically these geometrical

3 shapes are meshes (a generalization of a multi-dimensional cube, think grid-system) or tori (ring-layouts). In their 2003 paper, Yoshinaga et al. describe an algorithm for just such regular topographies as meshes and tori (1). Yoshinaga and his cohorts endeavored to create a router which was highly fault tolerant, and came up with a corresponding algorithm for it, which they titled Detour-NF, which could switch from deterministic routing to adaptive, and back, depending on a given situation and user preferences. Furthermore, they sought to expand the capabilities of the adaptive portion of their algorithm by providing a context in which both minimal and non-minimal path-selections could be available: Adaptive routing algorithms can be classified into two categories: minimal and non-minimal path selections [4]. Minimal adaptive algorithms are usually used to increase performance by improving the utilization of network resources, such as channels. However, they don t provide fault-tolerance. Therefore, non-minimal adaptive routing is required for reliable communication against faults. Once a non-minimal adaptive routing algorithm is supported, it is not so difficult to use the router as a minimal router by restricting path selection to the minimal paths. Design and Evaluation of a Fault-Tolerant Adaptive Router for Parallel Computers (Yoshinaga 1) In short, they aimed to create a fault-tolerant router that could take advantage of a wide spectrum of path-selection options with minimum necessary hardware and memory commission. They hoped to achieve this by taking any already-existing fully-adaptive routing algorithm, and changing its behavior with the use of some restrictions as well as freedoms. Fault-tolerance was added by taking the algorithm and adding one or more

4 VC s (in addition to any VC s the algorithm may already require) to any existing channels, which are treated as non-minimal-channels, as well as a function which details how to deal with these new VC s. The original algorithm and said function are then combined so that the new algorithm favors paths chosen by the original algorithm, but misroutes a message to a non-minimal VC channel if there is an unavoidable block in the path chosen by the original algorithm and no alternate minimal-routes. When a message s channel is blocked for some reason, the algorithm forces it to take a Westward, or negative turn, followed by any necessary turns, in order to leapfrog the faulty channel or blockage. Each leapfrog is started by heading in a negative direction, followed by appropriate positive or negative direction. Positive turns as the first step in a leapfrog, are not allowed. This is illustrated more clearly in their graphical representations, reproduced here. One unintended aspect of this algorithm is that if there is a blockage in the final turn of such a leapfrog, in a message headed in the Northerly, or positive direction, the route becomes unnecessarily longer, since a leapfrog, doesn t (Yoshinaga 3)

5 initiate from a positive position. Therefore they opted to allow 180 degree turns, also illustrated to the left, to allow a message to get to its destination quicker. The version of this algorithm for a torus structure fixes a weakness that does not allow the previous version (created for mesh-networks), to handle blockages on the negative edge. The torus algorithm overcomes this weakness by by utilizing wraparound channels that connect the nodes across the dateline. (3), by either preparing two of the non-minimal VC s in such a way as to eliminate torus cycles, or limiting the use of wraparound channels to a single hop per message. In essence, the torus algorithm ends up treating the network as a mesh anyway. Overall, though the hardware router meant to implement their algorithm is more expensive to build (a 29% increase in cost, measured in Flip-Flops, and a 56% increase in cell area), and the more accepted Duato s Protocol server does best in speed in each round, the Detour-NF routers have a much higher capacity for fault tolerance (which DP lacks), which gives it an advantage in random-traffic surge situations (5-7). Unfortunately, we don t get the chance to see the Detour-NF s fault-tolerance capabilities go up against other fault-tolerant routing algorithms, only against a Duato s run without any faults (which bests it every time), so we can t really get an answer as to it s comparative worth. The Detour-NF router is an admirable goal, which promised to see an increase in efficiency in any given situation due to said router s inherent pliability. A user could determine just what kind of message-routing algorithms would be used by his or her high-performance job, and presumably pick the most efficient options. Unfortunately, it

6 was an algorithm created for their specific router, and despite its potential for swappablealgorithms, it cannot be fully applied to a wider range of problems, unless the router comes with the deal, and with its marginal speedup in MHZ against a Duato s Protocol run, it may not be worth the investment. Another, more general, adaptive algorithm for regular mesh topographies comes from Ge-Ming Chiu of the IEEE Computer Society. His algorithm, the Odd-Even Turn Model, which functions without Virtual Channels, promises to greatly increase speedup compared to typical deadlock-avoiding, non-vc based algorithms, while sacrificing none of the deadlock protection. Chiu attempts to solve the un-evenness of previous non-vc, deadlock avoiding algorithms like the xy, turn model, west-first, north-last, and negativefirst algorithms, which work by either prohibiting certain turns a message can make, closely controlling them, or limiting their numbers in a general scope. Chiu redresses this unevenness by prohibiting very specific turns only in very specific situations, based on the current location. (1-2). Picture a mesh network as a grid, and label the top of the network North, the left West, the right East, and the bottom South. Then start labeling each column and row, left to right, top to bottom respectively, starting with an index of 0. Now consider North-West (NW) turn to be one where a message is traveling in a northern direction, then turns on a node, and begins traveling in a western direction. With this basic setup, you re ready to understand Chiu s algorithm, which consists of two basic rules: Rule 1. Any packet is not allowed to take an EN turn at any nodes located in an even column, and it is not allowed to take an NW turn at any nodes located in an odd column. Rule 2. Any packet is not allowed to take an ES turn at any

7 nodes located in an even column, and it is not allowed to take an SW turn at any nodes located in an odd column. The Odd-Even Turn Model for Adaptive Routing (Chiu 2) The basic premise behind these rules is that two messages taking mirror-opposite turns (for example, a message that s traveling East-to-South, and another traveling Eastto-North, but one row down), won t end up running into each other in the same column and causing a deadlock, a concept Chiu illustrated in his paper, reproduced below (3). Technically, the designations of even and odd in those rules can be interchanged, thus making the algorithm even more generally applicable. (Chiu3) The one weakness he acknowledges in his algorithm is that 180-degree turns cannot be allowed, lest it undermine the entire deadlock-avoidance system, standing in contrast, once again, to Detour-NF s insistence that they be allowed. The Odd-Even Turn Model, at the end, is just a small part of what makes a complete routing algorithm. But Chiu insists that with his turning-rules, it is easy to construct a fully-fledged routing algorithm that remains deadlock free, without the use of virtual-channels, by letting the turn-model give an adaptive algorithm options of where to

8 send the message for each hop Chiu provides such an algorithm, called ROUTE, a minimal-route algorithm. (3) Under Chiu s test, the odd-even turn model based algorithm performed well under high traffic conditions, and even handled hot-spot traffic tests the best. While the xy algorithm performed best in most simulations, its un-evenness and inadaptability proved to cause fluctuations in the networks performance, while the odd-even model held up reasonably well. The Odd-Even Turn Model stands in sharp contrast to the Detour-NF approach to routing. Where Detour-NF leaves room for third-party algorithms to potentially be swapped in, the Odd-Even Model is essentially itself a swappable portion of a larger algorithm. Where Detour seeks to limit routing, Odd-Even seeks to allow as much choice as possible, and where Detour insists on allowing a turn (namely a 180 degree turn), Odd- Even forbids it. Perhaps the most striking difference is Odd-Even s lack of Virtual Channels, which Detour-NF is dependent on. Citing the high-price of implementing VC s, Chiu determines that non-vc based, deadlock-free algorithms are important to produce. (1) The biggest advantage of Chiu s model is that it has the potential to be scaled and adapted for a variety of different and useful situations. The model itself, while providing a big piece of a deadlock-free routing algorithm, is also just a single part of one, allowing it to be integrated into more complex, or more streamlined and efficient proper routing algorithms. When discussing the importance of creating a non-vc-based routing algorithm, he points out that this very model could be used in other algorithms as a way

9 to manage VC s themselves. In addition, the model readily extends to higher-dimensional meshes, allowing it to be adapted for bigger and more expansive parallel highperformance environments. Now that we ve seen some algorithms for regular network topology, let s focus our attention on irregular topology. Where regular topology allows a network to be easily mapped to a set geometric shape or graph, irregular simply refuses to. A good example of an irregular network is the Internet, and its vast, uncounted, and difficult-totopographically-analyze structure. Let s take a look some routing algorithms that attempt to tackle this difficult topography. The Koibuchi article discusses algorithms for SANs with virtual channels and without virtual channels and compares them to show the advantages of SANs with virtual channels.(koibuchi 1) The irregular-topology algorithms we will be looking at fall into two basic categories. Low port first, random, Sancho s and Low virtual-channel first can all be done without virtual channels. Low port first, random,, low virtual-channel first, along with high virtual-channel first, high physical-channel first, and low physical-channel first are also able to take advantage of virtual channels. The first algorithm is the random selection. This algorithm doesn t take into account network congestions so while it is easy to implement is not an optimal solution. The low port first method, when given more than one available paths simply selects the port with the smallest port unique identifier. The low port first method also doesn t take into account network congestion and is only a simplistic method for directing network traffic.

10 The Up*/Down* routing algorithm proposed by Sancho is a method is a way to avoid network congestion. Up*/Down* routing does not have virtual channels for communication. It starts by going through the network and labeling all the channels with a counter that is equal to the number of legal paths crossing its channel. When faced with two or more possible paths you then remove the path with the largest counter. Then the network is relabeled with new counters and you continue to remove channels until you are left with a single path. This algorithm takes O(n^2*diameter) where n is the number of switches.(119) This may take a while but the time will be made up for in the reduction of congestion. The following algorithms use the Descending Layers (DL) routing method proposed by the Koibuchi article. DL routing improves routing over Up*/Down* by using virtual channels. DL routing uses smaller sub networks that work individually the same as Up*/Down* routing. DL routing then establishes paths across each of the different sub networks in order to insure there is a path from all nodes in the network to all other nodes in the network. For networks with routers capable of virtual channels there are better algorithms than Sanchos. Algorithms such as high virtual-channel first, high physical-channel first, low virtual-channel first, low physical-channel first can be better for avoiding network congestion. High virtual-channel first is different from Sancho s algorithm in that it selects the virtual channel with the largest counter and removes it where Sancho s removed the

11 regular channel with the largest counter. This is basically a way to use Sancho s algorithm on a system with virtual channels. (4) High physical-channel first is different in that it selects the physical channel with the largest sum of counters from every virtual channel. This algorithm avoids bottlenecks of virtual and physical channels while high virtual-channel first only avoids virtual channel bottlenecks. Bottlenecks occur when too much network traffic is sent through a single channel. This requires some messages to have to wait for the channel to clear to send and will decrease the time it takes to send a message. Low virtual-channel first is a bit different. It takes the virtual channel with the smallest count and removes all other paths making the selected one the chosen path. Similarly, low physical-channel first selects the smallest counter value on the physical channel based on the sum of the counters on each virtual channel. It also then fixes the channel and uses that path for communication. Low virtual-channel first spreads out the work and used the virtual network to keep congestion down. Low physical-channel first not only spreads the work over virtual channels it also spreads it out over physical channels.(4) High virtual channel first shows the best traffic throughput. There is a 92% increase when comparing the throughput of simplistic routine of low port first and high virtual-channel first on a DC routing network. This shows that the high virtual-channel first can send more data of the network per clock cycle. If virtual channels aren t available then you can still get a throughput increase of 90.5% by using high physicalchannel first. (6-14)

12 Low virtual-channel first and low physical-channel first do not show the speedup of the high algorithms but they do still show a significant improvement over low port first and random. Low physical-channel first gives an increase of 68.6% increase when comparing to low port. Low virtual-channel first gives an increase of 67.4% when also comparing to low port. (6-14) All algorithm in the DL routing system have higher throughput than the Up*/Down* routing algorithms. They Up*/Down* algorithms have smaller standard deviations meaning they are more consistent but the overall throughput is so much higher for DL routing that they are still the obviously better choice. This is because there are fewer channels crossing paths in the DL routing leading to shorter path hops and the algorithms are pages on the number of hops. Throughput is not the only way to measure the effectiveness of the path selection algorithms. Latency is also a central issue in evaluating different path selection algorithms. The two algorithms with the lowest latency are high virtual-channel and high physical-channel first. This means high virtual-channel and high-physical-channel not only give you the highest throughput but they also have the lowest latency of the algorithm we are looking at. The Up*/Down* routing algorithms all have worse latencies than the DC algorithms. These conclusions may not be the same if these algorithms are applied to a regular network structure. All the tests used for this discussion of path selection were based on irregular network structures. The structures of these algorithms are made for irregular network structures so results on regular topology could be different. In general, though

13 Sancho s, high virtual-channel first and high physical-channel first all perform better than low physical-channel first and low virtual-channel first for irregular topologies. Though there doesn t seem to be any single algorithm that can always be counted on for optimum optimization of SAN communications, each of these articles and algorithms show how, depending on your regular traffic patterns, network layout, and communications needs, you can usually find a way to adapt one or more algorithms (and sometimes combine them) to achieve satisfactory optimization, and keep your parallel network running with a minimum of overhead.

14 Works Cited Tsutomu Yoshinaga, Hiroyuki Hosogoshi, Masahiro Sowa, University of Electro- Communications, Tokyo, Japan, Design and Evaluation of a Fault-Tolerant Adaptive Router for Parallel Computers, Innovative Architecture for Future Generation High-Performance Processors and Systems , 2003, Kauai, HI Ge-Ming Chiu, Member, IEEE Computer Society. The Odd-Even Turn Model for Adaptive Routing. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 11, NO. 7, JULY 2000, pp Michihiro Koibuchi *, Akiya Jouraku, Hideharu Amano. Department of Information and Computer Science, Keio University. Path selection algorithm: the strategy for designing deterministic routing from alternative paths. Parallel Computing 31 (2005) Networks and Topologies,

Design and Evaluation of a Fault-Tolerant Adaptive Router for Parallel Computers

Design and Evaluation of a Fault-Tolerant Adaptive Router for Parallel Computers Design and Evaluation of a Fault-Tolerant Adaptive Router for Parallel Computers Tsutomu YOSHINAGA, Hiroyuki HOSOGOSHI, Masahiro SOWA Graduate School of Information Systems, University of Electro-Communications,

More information

Interconnection topologies (cont.) [ ] In meshes and hypercubes, the average distance increases with the dth root of N.

Interconnection topologies (cont.) [ ] In meshes and hypercubes, the average distance increases with the dth root of N. Interconnection topologies (cont.) [ 10.4.4] In meshes and hypercubes, the average distance increases with the dth root of N. In a tree, the average distance grows only logarithmically. A simple tree structure,

More information

Fault-Tolerant Routing Algorithm in Meshes with Solid Faults

Fault-Tolerant Routing Algorithm in Meshes with Solid Faults Fault-Tolerant Routing Algorithm in Meshes with Solid Faults Jong-Hoon Youn Bella Bose Seungjin Park Dept. of Computer Science Dept. of Computer Science Dept. of Computer Science Oregon State University

More information

Recall: The Routing problem: Local decisions. Recall: Multidimensional Meshes and Tori. Properties of Routing Algorithms

Recall: The Routing problem: Local decisions. Recall: Multidimensional Meshes and Tori. Properties of Routing Algorithms CS252 Graduate Computer Architecture Lecture 16 Multiprocessor Networks (con t) March 14 th, 212 John Kubiatowicz Electrical Engineering and Computer Sciences University of California, Berkeley http://www.eecs.berkeley.edu/~kubitron/cs252

More information

Interconnection Networks: Routing. Prof. Natalie Enright Jerger

Interconnection Networks: Routing. Prof. Natalie Enright Jerger Interconnection Networks: Routing Prof. Natalie Enright Jerger Routing Overview Discussion of topologies assumed ideal routing In practice Routing algorithms are not ideal Goal: distribute traffic evenly

More information

Routing Algorithms. Review

Routing Algorithms. Review Routing Algorithms Today s topics: Deterministic, Oblivious Adaptive, & Adaptive models Problems: efficiency livelock deadlock 1 CS6810 Review Network properties are a combination topology topology dependent

More information

A Deterministic Fault-Tolerant and Deadlock-Free Routing Protocol in 2-D Meshes Based on Odd-Even Turn Model

A Deterministic Fault-Tolerant and Deadlock-Free Routing Protocol in 2-D Meshes Based on Odd-Even Turn Model A Deterministic Fault-Tolerant and Deadlock-Free Routing Protocol in 2-D Meshes Based on Odd-Even Turn Model Jie Wu Dept. of Computer Science and Engineering Florida Atlantic University Boca Raton, FL

More information

Fault-adaptive routing

Fault-adaptive routing Fault-adaptive routing Presenter: Zaheer Ahmed Supervisor: Adan Kohler Reviewers: Prof. Dr. M. Radetzki Prof. Dr. H.-J. Wunderlich Date: 30-June-2008 7/2/2009 Agenda Motivation Fundamentals of Routing

More information

Lecture 24: Interconnection Networks. Topics: topologies, routing, deadlocks, flow control

Lecture 24: Interconnection Networks. Topics: topologies, routing, deadlocks, flow control Lecture 24: Interconnection Networks Topics: topologies, routing, deadlocks, flow control 1 Topology Examples Grid Torus Hypercube Criteria Bus Ring 2Dtorus 6-cube Fully connected Performance Bisection

More information

Chapter 1.5 Data Transmission and Networking.

Chapter 1.5 Data Transmission and Networking. Chapter 1.5 Data Transmission and Networking. 1.5 (a) Networks All the systems that have been mentioned so far have been individual computers, sometimes with more than one user, but single processors.

More information

Lecture 13: Interconnection Networks. Topics: lots of background, recent innovations for power and performance

Lecture 13: Interconnection Networks. Topics: lots of background, recent innovations for power and performance Lecture 13: Interconnection Networks Topics: lots of background, recent innovations for power and performance 1 Interconnection Networks Recall: fully connected network, arrays/rings, meshes/tori, trees,

More information

Fault-Tolerant Wormhole Routing Algorithms in Meshes in the Presence of Concave Faults

Fault-Tolerant Wormhole Routing Algorithms in Meshes in the Presence of Concave Faults Fault-Tolerant Wormhole Routing Algorithms in Meshes in the Presence of Concave Faults Seungjin Park Jong-Hoon Youn Bella Bose Dept. of Computer Science Dept. of Computer Science Dept. of Computer Science

More information

From Routing to Traffic Engineering

From Routing to Traffic Engineering 1 From Routing to Traffic Engineering Robert Soulé Advanced Networking Fall 2016 2 In the beginning B Goal: pair-wise connectivity (get packets from A to B) Approach: configure static rules in routers

More information

Lecture 12: Interconnection Networks. Topics: communication latency, centralized and decentralized switches, routing, deadlocks (Appendix E)

Lecture 12: Interconnection Networks. Topics: communication latency, centralized and decentralized switches, routing, deadlocks (Appendix E) Lecture 12: Interconnection Networks Topics: communication latency, centralized and decentralized switches, routing, deadlocks (Appendix E) 1 Topologies Internet topologies are not very regular they grew

More information

Problem Formulation. Specialized algorithms are required for clock (and power nets) due to strict specifications for routing such nets.

Problem Formulation. Specialized algorithms are required for clock (and power nets) due to strict specifications for routing such nets. Clock Routing Problem Formulation Specialized algorithms are required for clock (and power nets) due to strict specifications for routing such nets. Better to develop specialized routers for these nets.

More information

NOW Handout Page 1. Outline. Networks: Routing and Design. Routing. Routing Mechanism. Routing Mechanism (cont) Properties of Routing Algorithms

NOW Handout Page 1. Outline. Networks: Routing and Design. Routing. Routing Mechanism. Routing Mechanism (cont) Properties of Routing Algorithms Outline Networks: Routing and Design Routing Switch Design Case Studies CS 5, Spring 99 David E. Culler Computer Science Division U.C. Berkeley 3/3/99 CS5 S99 Routing Recall: routing algorithm determines

More information

Data Communication. Introduction of Communication. Data Communication. Elements of Data Communication (Communication Model)

Data Communication. Introduction of Communication. Data Communication. Elements of Data Communication (Communication Model) Data Communication Introduction of Communication The need to communicate is part of man s inherent being. Since the beginning of time the human race has communicated using different techniques and methods.

More information

Fault-Tolerant and Deadlock-Free Routing in 2-D Meshes Using Rectilinear-Monotone Polygonal Fault Blocks

Fault-Tolerant and Deadlock-Free Routing in 2-D Meshes Using Rectilinear-Monotone Polygonal Fault Blocks Fault-Tolerant and Deadlock-Free Routing in -D Meshes Using Rectilinear-Monotone Polygonal Fault Blocks Jie Wu Department of Computer Science and Engineering Florida Atlantic University Boca Raton, FL

More information

NOC Deadlock and Livelock

NOC Deadlock and Livelock NOC Deadlock and Livelock 1 Deadlock (When?) Deadlock can occur in an interconnection network, when a group of packets cannot make progress, because they are waiting on each other to release resource (buffers,

More information

ECE 669 Parallel Computer Architecture

ECE 669 Parallel Computer Architecture ECE 669 Parallel Computer Architecture Lecture 21 Routing Outline Routing Switch Design Flow Control Case Studies Routing Routing algorithm determines which of the possible paths are used as routes how

More information

Interconnection Networks: Topology. Prof. Natalie Enright Jerger

Interconnection Networks: Topology. Prof. Natalie Enright Jerger Interconnection Networks: Topology Prof. Natalie Enright Jerger Topology Overview Definition: determines arrangement of channels and nodes in network Analogous to road map Often first step in network design

More information

EE 382C Interconnection Networks

EE 382C Interconnection Networks EE 8C Interconnection Networks Deadlock and Livelock Stanford University - EE8C - Spring 6 Deadlock and Livelock: Terminology Deadlock: A condition in which an agent waits indefinitely trying to acquire

More information

1 GSW Bridging and Switching

1 GSW Bridging and Switching 1 Sandwiched between the physical and media access layers of local area networking (such as Ethernet) and the routeing of the Internet layer of the IP protocol, lies the thorny subject of bridges. Bridges

More information

Deadlock and Livelock. Maurizio Palesi

Deadlock and Livelock. Maurizio Palesi Deadlock and Livelock 1 Deadlock (When?) Deadlock can occur in an interconnection network, when a group of packets cannot make progress, because they are waiting on each other to release resource (buffers,

More information

McGill University - Faculty of Engineering Department of Electrical and Computer Engineering

McGill University - Faculty of Engineering Department of Electrical and Computer Engineering McGill University - Faculty of Engineering Department of Electrical and Computer Engineering ECSE 494 Telecommunication Networks Lab Prof. M. Coates Winter 2003 Experiment 5: LAN Operation, Multiple Access

More information

Performance Analysis of NoC Architectures

Performance Analysis of NoC Architectures Performance Analysis of NoC Architectures Anitha G #1,Muralidharan D *2,Muthaiah R #3 School of Computing, SASTRA University,Thanjavur,India #1 mailtoanitharaj@gmail.com, *2 murali@core.sastra.edu, #3

More information

Fault-Tolerant Routing in Fault Blocks. Planarly Constructed. Dong Xiang, Jia-Guang Sun, Jie. and Krishnaiyan Thulasiraman. Abstract.

Fault-Tolerant Routing in Fault Blocks. Planarly Constructed. Dong Xiang, Jia-Guang Sun, Jie. and Krishnaiyan Thulasiraman. Abstract. Fault-Tolerant Routing in Fault Blocks Planarly Constructed Dong Xiang, Jia-Guang Sun, Jie and Krishnaiyan Thulasiraman Abstract A few faulty nodes can an n-dimensional mesh or torus network unsafe for

More information

Network-on-chip (NOC) Topologies

Network-on-chip (NOC) Topologies Network-on-chip (NOC) Topologies 1 Network Topology Static arrangement of channels and nodes in an interconnection network The roads over which packets travel Topology chosen based on cost and performance

More information

BARP-A Dynamic Routing Protocol for Balanced Distribution of Traffic in NoCs

BARP-A Dynamic Routing Protocol for Balanced Distribution of Traffic in NoCs -A Dynamic Routing Protocol for Balanced Distribution of Traffic in NoCs Pejman Lotfi-Kamran, Masoud Daneshtalab *, Caro Lucas, and Zainalabedin Navabi School of Electrical and Computer Engineering, The

More information

The Odd-Even Turn Model for Adaptive Routing

The Odd-Even Turn Model for Adaptive Routing IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 11, NO. 7, JULY 2000 729 The Odd-Even Turn Model for Adaptive Routing Ge-Ming Chiu, Member, IEEE Computer Society AbstractÐThis paper presents

More information

Lecture 12: Interconnection Networks. Topics: dimension/arity, routing, deadlock, flow control

Lecture 12: Interconnection Networks. Topics: dimension/arity, routing, deadlock, flow control Lecture 12: Interconnection Networks Topics: dimension/arity, routing, deadlock, flow control 1 Interconnection Networks Recall: fully connected network, arrays/rings, meshes/tori, trees, butterflies,

More information

Media-Ready Network Transcript

Media-Ready Network Transcript Media-Ready Network Transcript Hello and welcome to this Cisco on Cisco Seminar. I m Bob Scarbrough, Cisco IT manager on the Cisco on Cisco team. With me today are Sheila Jordan, Vice President of the

More information

Interconnection Network

Interconnection Network Interconnection Network Recap: Generic Parallel Architecture A generic modern multiprocessor Network Mem Communication assist (CA) $ P Node: processor(s), memory system, plus communication assist Network

More information

Communication in Multicomputers with Nonconvex Faults

Communication in Multicomputers with Nonconvex Faults Communication in Multicomputers with Nonconvex Faults Suresh Chalasani Rajendra V. Boppana Technical Report : CS-96-12 October 1996 The University of Texas at San Antonio Division of Computer Science San

More information

Missing Pieces of the Puzzle

Missing Pieces of the Puzzle Missing Pieces of the Puzzle EE122 Fall 2011 Scott Shenker http://inst.eecs.berkeley.edu/~ee122/ Materials with thanks to Jennifer Rexford, Ion Stoica, Vern Paxson and other colleagues at Princeton and

More information

Resource Guide Implementing QoS for WX/WXC Application Acceleration Platforms

Resource Guide Implementing QoS for WX/WXC Application Acceleration Platforms Resource Guide Implementing QoS for WX/WXC Application Acceleration Platforms Juniper Networks, Inc. 1194 North Mathilda Avenue Sunnyvale, CA 94089 USA 408 745 2000 or 888 JUNIPER www.juniper.net Table

More information

OFAR-CM: Efficient Dragonfly Networks with Simple Congestion Management

OFAR-CM: Efficient Dragonfly Networks with Simple Congestion Management Marina Garcia 22 August 2013 OFAR-CM: Efficient Dragonfly Networks with Simple Congestion Management M. Garcia, E. Vallejo, R. Beivide, M. Valero and G. Rodríguez Document number OFAR-CM: Efficient Dragonfly

More information

The Drunken Sailor s Challenge

The Drunken Sailor s Challenge The Drunken Sailor s Challenge CS7001 mini-project by Arya Irani with Tucker Balch September 29, 2003 Introduction The idea [1] is to design an agent to navigate a two-dimensional field of obstacles given

More information

GUIDE. Optimal Network Designs with Cohesity

GUIDE. Optimal Network Designs with Cohesity Optimal Network Designs with Cohesity TABLE OF CONTENTS Introduction...3 Key Concepts...4 Five Common Configurations...5 3.1 Simple Topology...5 3.2 Standard Topology...6 3.3 Layered Topology...7 3.4 Cisco

More information

Interconnect Technology and Computational Speed

Interconnect Technology and Computational Speed Interconnect Technology and Computational Speed From Chapter 1 of B. Wilkinson et al., PARAL- LEL PROGRAMMING. Techniques and Applications Using Networked Workstations and Parallel Computers, augmented

More information

Making Sense of Home Wi-Fi Coverage Solutions

Making Sense of Home Wi-Fi Coverage Solutions Making Sense of Home Wi-Fi Coverage Solutions www.xcellair.com Phone: +1 858.412.0186 Email: sales@xcellair.com 6540 Lusk Blvd, Suite C102, San Diego, CA 92121 Introduction Wireless connectivity is so

More information

Lecture 15: PCM, Networks. Today: PCM wrap-up, projects discussion, on-chip networks background

Lecture 15: PCM, Networks. Today: PCM wrap-up, projects discussion, on-chip networks background Lecture 15: PCM, Networks Today: PCM wrap-up, projects discussion, on-chip networks background 1 Hard Error Tolerance in PCM PCM cells will eventually fail; important to cause gradual capacity degradation

More information

Year. Small Steps Guidance and Examples. Block 1 Properties of Shapes. Released March 2018

Year. Small Steps Guidance and Examples. Block 1 Properties of Shapes. Released March 2018 Released March 2018 The sequence of small steps has been produced by White Rose Maths. White Rose Maths gives permission to schools and teachers to use the small steps in their own teaching in their own

More information

PDA-HyPAR: Path-Diversity-Aware Hybrid Planar Adaptive Routing Algorithm for 3D NoCs

PDA-HyPAR: Path-Diversity-Aware Hybrid Planar Adaptive Routing Algorithm for 3D NoCs PDA-HyPAR: Path-Diversity-Aware Hybrid Planar Adaptive Routing Algorithm for 3D NoCs Jindun Dai *1,2, Renjie Li 2, Xin Jiang 3, Takahiro Watanabe 2 1 Department of Electrical Engineering, Shanghai Jiao

More information

Welfare Navigation Using Genetic Algorithm

Welfare Navigation Using Genetic Algorithm Welfare Navigation Using Genetic Algorithm David Erukhimovich and Yoel Zeldes Hebrew University of Jerusalem AI course final project Abstract Using standard navigation algorithms and applications (such

More information

Basic Switch Organization

Basic Switch Organization NOC Routing 1 Basic Switch Organization 2 Basic Switch Organization Link Controller Used for coordinating the flow of messages across the physical link of two adjacent switches 3 Basic Switch Organization

More information

Prediction Router: Yet another low-latency on-chip router architecture

Prediction Router: Yet another low-latency on-chip router architecture Prediction Router: Yet another low-latency on-chip router architecture Hiroki Matsutani Michihiro Koibuchi Hideharu Amano Tsutomu Yoshinaga (Keio Univ., Japan) (NII, Japan) (Keio Univ., Japan) (UEC, Japan)

More information

(Refer Slide Time: 00:01:30)

(Refer Slide Time: 00:01:30) Digital Circuits and Systems Prof. S. Srinivasan Department of Electrical Engineering Indian Institute of Technology, Madras Lecture - 32 Design using Programmable Logic Devices (Refer Slide Time: 00:01:30)

More information

Topologies. Maurizio Palesi. Maurizio Palesi 1

Topologies. Maurizio Palesi. Maurizio Palesi 1 Topologies Maurizio Palesi Maurizio Palesi 1 Network Topology Static arrangement of channels and nodes in an interconnection network The roads over which packets travel Topology chosen based on cost and

More information

CSC 2515 Introduction to Machine Learning Assignment 2

CSC 2515 Introduction to Machine Learning Assignment 2 CSC 2515 Introduction to Machine Learning Assignment 2 Zhongtian Qiu(1002274530) Problem 1 See attached scan files for question 1. 2. Neural Network 2.1 Examine the statistics and plots of training error

More information

Interprocessor Communication. Basics of Network Routing

Interprocessor Communication. Basics of Network Routing Interprocessor Communication There are two main differences between sequential computers and parallel computers -- multiple processors and the hardware to connect them together. That hardware is the most

More information

Specifying Storage Servers for IP security applications

Specifying Storage Servers for IP security applications Specifying Storage Servers for IP security applications The migration of security systems from analogue to digital IP based solutions has created a large demand for storage servers high performance PCs

More information

JUNCTION BASED ROUTING: A NOVEL TECHNIQUE FOR LARGE NETWORK ON CHIP PLATFORMS

JUNCTION BASED ROUTING: A NOVEL TECHNIQUE FOR LARGE NETWORK ON CHIP PLATFORMS 1 JUNCTION BASED ROUTING: A NOVEL TECHNIQUE FOR LARGE NETWORK ON CHIP PLATFORMS Shabnam Badri THESIS WORK 2011 ELECTRONICS JUNCTION BASED ROUTING: A NOVEL TECHNIQUE FOR LARGE NETWORK ON CHIP PLATFORMS

More information

(Refer Slide Time: 00:02:00)

(Refer Slide Time: 00:02:00) Computer Graphics Prof. Sukhendu Das Dept. of Computer Science and Engineering Indian Institute of Technology, Madras Lecture - 18 Polyfill - Scan Conversion of a Polygon Today we will discuss the concepts

More information

SOFTWARE BASED FAULT-TOLERANT OBLIVIOUS ROUTING IN PIPELINED NETWORKS*

SOFTWARE BASED FAULT-TOLERANT OBLIVIOUS ROUTING IN PIPELINED NETWORKS* SOFTWARE BASED FAULT-TOLERANT OBLIVIOUS ROUTING IN PIPELINED NETWORKS* Young-Joo Suh, Binh Vien Dao, Jose Duato, and Sudhakar Yalamanchili Computer Systems Research Laboratory Facultad de Informatica School

More information

Lecture: Interconnection Networks

Lecture: Interconnection Networks Lecture: Interconnection Networks Topics: Router microarchitecture, topologies Final exam next Tuesday: same rules as the first midterm 1 Packets/Flits A message is broken into multiple packets (each packet

More information

INTERCONNECTION NETWORKS LECTURE 4

INTERCONNECTION NETWORKS LECTURE 4 INTERCONNECTION NETWORKS LECTURE 4 DR. SAMMAN H. AMEEN 1 Topology Specifies way switches are wired Affects routing, reliability, throughput, latency, building ease Routing How does a message get from source

More information

Lecture: Interconnection Networks. Topics: TM wrap-up, routing, deadlock, flow control, virtual channels

Lecture: Interconnection Networks. Topics: TM wrap-up, routing, deadlock, flow control, virtual channels Lecture: Interconnection Networks Topics: TM wrap-up, routing, deadlock, flow control, virtual channels 1 TM wrap-up Eager versioning: create a log of old values Handling problematic situations with a

More information

This chapter provides the background knowledge about Multistage. multistage interconnection networks are explained. The need, objectives, research

This chapter provides the background knowledge about Multistage. multistage interconnection networks are explained. The need, objectives, research CHAPTER 1 Introduction This chapter provides the background knowledge about Multistage Interconnection Networks. Metrics used for measuring the performance of various multistage interconnection networks

More information

Distributed Sorting. Chapter Array & Mesh

Distributed Sorting. Chapter Array & Mesh Chapter 9 Distributed Sorting Indeed, I believe that virtually every important aspect of programming arises somewhere in the context of sorting [and searching]! Donald E. Knuth, The Art of Computer Programming

More information

ECE 1749H: Interconnec1on Networks for Parallel Computer Architectures: Rou1ng. Prof. Natalie Enright Jerger

ECE 1749H: Interconnec1on Networks for Parallel Computer Architectures: Rou1ng. Prof. Natalie Enright Jerger ECE 1749H: Interconnec1on Networks for Parallel Computer Architectures: Rou1ng Prof. Natalie Enright Jerger Announcements Feedback on your project proposals This week Scheduled extended 1 week Next week:

More information

CS519: Computer Networks. Lecture 2: Feb 2, 2004 IP (Internet Protocol)

CS519: Computer Networks. Lecture 2: Feb 2, 2004 IP (Internet Protocol) : Computer Networks Lecture 2: Feb 2, 2004 IP (Internet Protocol) A hypothetical service You want a mail delivery service You have two choices: Acme Guaranteed Mail Delivery Service We never fail Rocko

More information

TDT Appendix E Interconnection Networks

TDT Appendix E Interconnection Networks TDT 4260 Appendix E Interconnection Networks Review Advantages of a snooping coherency protocol? Disadvantages of a snooping coherency protocol? Advantages of a directory coherency protocol? Disadvantages

More information

EECS 578 Interconnect Mini-project

EECS 578 Interconnect Mini-project EECS578 Bertacco Fall 2015 EECS 578 Interconnect Mini-project Assigned 09/17/15 (Thu) Due 10/02/15 (Fri) Introduction In this mini-project, you are asked to answer questions about issues relating to interconnect

More information

Memory. From Chapter 3 of High Performance Computing. c R. Leduc

Memory. From Chapter 3 of High Performance Computing. c R. Leduc Memory From Chapter 3 of High Performance Computing c 2002-2004 R. Leduc Memory Even if CPU is infinitely fast, still need to read/write data to memory. Speed of memory increasing much slower than processor

More information

MESH-CONNECTED networks have been widely used in

MESH-CONNECTED networks have been widely used in 620 IEEE TRANSACTIONS ON COMPUTERS, VOL. 58, NO. 5, MAY 2009 Practical Deadlock-Free Fault-Tolerant Routing in Meshes Based on the Planar Network Fault Model Dong Xiang, Senior Member, IEEE, Yueli Zhang,

More information

On Constructing the Minimum Orthogonal Convex Polygon in 2-D Faulty Meshes

On Constructing the Minimum Orthogonal Convex Polygon in 2-D Faulty Meshes On Constructing the Minimum Orthogonal Convex Polygon in 2-D Faulty Meshes Jie Wu Department of Computer Science and Engineering Florida Atlantic University Boca Raton, FL 33431 E-mail: jie@cse.fau.edu

More information

Deadlock and Router Micro-Architecture

Deadlock and Router Micro-Architecture 1 EE482: Advanced Computer Organization Lecture #8 Interconnection Network Architecture and Design Stanford University 22 April 1999 Deadlock and Router Micro-Architecture Lecture #8: 22 April 1999 Lecturer:

More information

THE COMPLETE FIELD GUIDE TO THE WAN

THE COMPLETE FIELD GUIDE TO THE WAN THE COMPLETE FIELD GUIDE TO THE WAN People interested in setting up a wide-area network (WAN) often have very specific reasons for wanting to do so. WANs are specialized technological tools that deliver

More information

CS 162 Operating Systems and Systems Programming Professor: Anthony D. Joseph Spring Lecture 20: Networks and Distributed Systems

CS 162 Operating Systems and Systems Programming Professor: Anthony D. Joseph Spring Lecture 20: Networks and Distributed Systems S 162 Operating Systems and Systems Programming Professor: Anthony D. Joseph Spring 2003 Lecture 20: Networks and Distributed Systems 20.0 Main Points Motivation for distributed vs. centralized systems

More information

Raid: Who What Where When and Why. 3/23 Draft V1 David Arts 4/3 Draft V2 David Arts 4/10 First Release David Arts

Raid: Who What Where When and Why. 3/23 Draft V1 David Arts 4/3 Draft V2 David Arts 4/10 First Release David Arts Raid: Who What Where When and Why 3/23 Draft V1 David Arts 4/3 Draft V2 David Arts 4/10 First Release David Arts 1 Table of Contents General Concepts and Definitions... 3 What is Raid... 3 Origins of RAID...

More information

On Constructing the Minimum Orthogonal Convex Polygon for the Fault-Tolerant Routing in 2-D Faulty Meshes 1

On Constructing the Minimum Orthogonal Convex Polygon for the Fault-Tolerant Routing in 2-D Faulty Meshes 1 On Constructing the Minimum Orthogonal Convex Polygon for the Fault-Tolerant Routing in 2-D Faulty Meshes 1 Jie Wu Department of Computer Science and Engineering Florida Atlantic University Boca Raton,

More information

Dynamic Stress Wormhole Routing for Spidergon NoC with effective fault tolerance and load distribution

Dynamic Stress Wormhole Routing for Spidergon NoC with effective fault tolerance and load distribution Dynamic Stress Wormhole Routing for Spidergon NoC with effective fault tolerance and load distribution Nishant Satya Lakshmikanth sailtosatya@gmail.com Krishna Kumaar N.I. nikrishnaa@gmail.com Sudha S

More information

Assignment 5. Georgia Koloniari

Assignment 5. Georgia Koloniari Assignment 5 Georgia Koloniari 2. "Peer-to-Peer Computing" 1. What is the definition of a p2p system given by the authors in sec 1? Compare it with at least one of the definitions surveyed in the last

More information

A Simple Placement and Routing Algorithm for a Two-Dimensional Computational Origami Architecture

A Simple Placement and Routing Algorithm for a Two-Dimensional Computational Origami Architecture A Simple Placement and Routing Algorithm for a Two-Dimensional Computational Origami Architecture Robert S. French April 5, 1989 Abstract Computational origami is a parallel-processing concept in which

More information

6.033 Spring 2015 Lecture #11: Transport Layer Congestion Control Hari Balakrishnan Scribed by Qian Long

6.033 Spring 2015 Lecture #11: Transport Layer Congestion Control Hari Balakrishnan Scribed by Qian Long 6.033 Spring 2015 Lecture #11: Transport Layer Congestion Control Hari Balakrishnan Scribed by Qian Long Please read Chapter 19 of the 6.02 book for background, especially on acknowledgments (ACKs), timers,

More information

Routing and Deadlock

Routing and Deadlock 3.5-1 3.5-1 Routing and Deadlock Routing would be easy...... were it not for possible deadlock. Topics For This Set: Routing definitions. Deadlock definitions. Resource dependencies. Acyclic deadlock free

More information

NETWORK-BASED parallel processing using system area

NETWORK-BASED parallel processing using system area 320 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 18, NO. 3, MARCH 2007 An Effective Design of Deadlock-Free Routing Algorithms Based on 2D Turn Model for Irregular Networks Akiya Jouraku,

More information

CS125 : Introduction to Computer Science. Lecture Notes #38 and #39 Quicksort. c 2005, 2003, 2002, 2000 Jason Zych

CS125 : Introduction to Computer Science. Lecture Notes #38 and #39 Quicksort. c 2005, 2003, 2002, 2000 Jason Zych CS125 : Introduction to Computer Science Lecture Notes #38 and #39 Quicksort c 2005, 2003, 2002, 2000 Jason Zych 1 Lectures 38 and 39 : Quicksort Quicksort is the best sorting algorithm known which is

More information

Learning the Binary System

Learning the Binary System Learning the Binary System www.brainlubeonline.com/counting_on_binary/ Formated to L A TEX: /25/22 Abstract This is a document on the base-2 abstract numerical system, or Binary system. This is a VERY

More information

1 of 6 4/8/2011 4:08 PM Electronic Hardware Information, Guides and Tools search newsletter subscribe Home Utilities Downloads Links Info Ads by Google Raid Hard Drives Raid Raid Data Recovery SSD in Raid

More information

CS BGP v4. Fall 2014

CS BGP v4. Fall 2014 CS 457 - BGP v4 Fall 2014 Autonomous Systems What is an AS? a set of routers under a single technical administration uses an interior gateway protocol (IGP) and common metrics to route packets within the

More information

Pathfinding Algorithms and Implementations on Grid Map

Pathfinding Algorithms and Implementations on Grid Map Pathfinding Algorithms and Implementations on Grid Map Steven Andrew / 13509061 Program Studi Teknik Informatika Sekolah Teknik Elektro dan Informatika Institut Teknologi Bandung, Jl. Ganesha 10 Bandung

More information

A Case for Random Shortcut Topologies for HPC Interconnects

A Case for Random Shortcut Topologies for HPC Interconnects A Case for Random Shortcut Topologies for HPC Interconnects Michihiro Koibuchi National Institute of Informatics / SOKENDAI 2-1-2, Hitotsubashi, Chiyoda-ku, Tokyo, JAPAN 11-843 koibuchi@nii.ac.jp Hiroki

More information

Growth. Individual departments in a university buy LANs for their own machines and eventually want to interconnect with other campus LANs.

Growth. Individual departments in a university buy LANs for their own machines and eventually want to interconnect with other campus LANs. Internetworking Multiple networks are a fact of life: Growth. Individual departments in a university buy LANs for their own machines and eventually want to interconnect with other campus LANs. Fault isolation,

More information

CMSC 611: Advanced. Interconnection Networks

CMSC 611: Advanced. Interconnection Networks CMSC 611: Advanced Computer Architecture Interconnection Networks Interconnection Networks Massively parallel processor networks (MPP) Thousands of nodes Short distance (

More information

Assignment 7: TCP and Congestion Control Due the week of October 29/30, 2015

Assignment 7: TCP and Congestion Control Due the week of October 29/30, 2015 Assignment 7: TCP and Congestion Control Due the week of October 29/30, 2015 I d like to complete our exploration of TCP by taking a close look at the topic of congestion control in TCP. To prepare for

More information

/633 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Priority Queues / Heaps Date: 9/27/17

/633 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Priority Queues / Heaps Date: 9/27/17 01.433/33 Introduction to Algorithms Lecturer: Michael Dinitz Topic: Priority Queues / Heaps Date: 9/2/1.1 Introduction In this lecture we ll talk about a useful abstraction, priority queues, which are

More information

Lecture 6: Overlay Networks. CS 598: Advanced Internetworking Matthew Caesar February 15, 2011

Lecture 6: Overlay Networks. CS 598: Advanced Internetworking Matthew Caesar February 15, 2011 Lecture 6: Overlay Networks CS 598: Advanced Internetworking Matthew Caesar February 15, 2011 1 Overlay networks: Motivations Protocol changes in the network happen very slowly Why? Internet is shared

More information

Lecture 16: On-Chip Networks. Topics: Cache networks, NoC basics

Lecture 16: On-Chip Networks. Topics: Cache networks, NoC basics Lecture 16: On-Chip Networks Topics: Cache networks, NoC basics 1 Traditional Networks Huh et al. ICS 05, Beckmann MICRO 04 Example designs for contiguous L2 cache regions 2 Explorations for Optimality

More information

You will begin by exploring the locations of the long term care facilities in Massachusetts using descriptive statistics.

You will begin by exploring the locations of the long term care facilities in Massachusetts using descriptive statistics. Getting Started 1. Create a folder on the desktop and call it your last name. 2. Copy and paste the data you will need to your folder from the folder specified by the instructor. Exercise 1: Explore the

More information

Comparing Implementations of Optimal Binary Search Trees

Comparing Implementations of Optimal Binary Search Trees Introduction Comparing Implementations of Optimal Binary Search Trees Corianna Jacoby and Alex King Tufts University May 2017 In this paper we sought to put together a practical comparison of the optimality

More information

IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 1, JANUARY

IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 1, JANUARY IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 1, JANUARY 2014 113 ZoneDefense: A Fault-Tolerant Routing for 2-D Meshes Without Virtual Channels Binzhang Fu, Member, IEEE,

More information

Data parallel algorithms 1

Data parallel algorithms 1 Data parallel algorithms (Guy Steele): The data-parallel programming style is an approach to organizing programs suitable for execution on massively parallel computers. In this lecture, we will characterize

More information

Mladen Stefanov F48235 R.A.I.D

Mladen Stefanov F48235 R.A.I.D R.A.I.D Data is the most valuable asset of any business today. Lost data, in most cases, means lost business. Even if you backup regularly, you need a fail-safe way to ensure that your data is protected

More information

Extending the LAN. Context. Info 341 Networking and Distributed Applications. Building up the network. How to hook things together. Media NIC 10/18/10

Extending the LAN. Context. Info 341 Networking and Distributed Applications. Building up the network. How to hook things together. Media NIC 10/18/10 Extending the LAN Info 341 Networking and Distributed Applications Context Building up the network Media NIC Application How to hook things together Transport Internetwork Network Access Physical Internet

More information

Slashing Downtime from 24 Hours to 24 Minutes:

Slashing Downtime from 24 Hours to 24 Minutes: Slashing Downtime from 24 Hours to 24 Minutes: Technology Advancements Make Warm-Site Disaster Recovery Possible Don Beyer Director, Technical Services Oakwood Healthcare Louie Caschera Chief Information

More information

The evasive speed of your Internet

The evasive speed of your Internet The evasive speed of your Internet 2018 Charter Communications. All rights reserved. Not all products, pricing and services are available in all areas. Pricing and actual speeds may vary. Restrictions

More information

You may use a calculator for these practice questions. You may

You may use a calculator for these practice questions. You may 660 Math Smart Practice Questions You may use a calculator for these practice questions. You may not know all the math to complete these practice questions yet, but try to think them through! 1. Eric lives

More information

MIDTERM EXAMINATION Networked Life (NETS 112) November 21, 2013 Prof. Michael Kearns

MIDTERM EXAMINATION Networked Life (NETS 112) November 21, 2013 Prof. Michael Kearns MIDTERM EXAMINATION Networked Life (NETS 112) November 21, 2013 Prof. Michael Kearns This is a closed-book exam. You should have no material on your desk other than the exam itself and a pencil or pen.

More information