A fundamental problem that confronts peer-to-peer applications is to efficiently locate the node that stores a particular data item. This paper presents Chord, a distributed lookup protocol that addresses this problem. Chord provides support for just one operation: given a key, it maps the key onto a node. Data location can be easily implemented on top of Chord by associating a key with each data item, and storing the key/data item pair at the node to which the key maps. Chord adapts efficiently as nodes join and leave the system, and can answer queries even if the system is continuously changing. Results from theoretical analysis, simulations, and experiments show that Chord is scalable, with communication cost and the state maintained by each node scaling logarithmically with the number of Chord nodes.

Chord: A scalable peer-to-peer lookup service for internet applications

Recently, a number of researchers have investigated a class of graph partitioning algorithms that reduce the size of the graph by collapsing vertices and edges, partition the smaller graph, and then uncoarsen it to construct a partition for the original graph [Bui and Jones, Proc. of the 6th SIAM Conference on Parallel Processing for Scientific Computing, 1993, 445--452; Hendrickson and Leland, A Multilevel Algorithm for Partitioning Graphs, Tech. report SAND 93-1301, Sandia National Laboratories, Albuquerque, NM, 1993]. From the early work it was clear that multilevel techniques held great promise; however, it was not known if they can be made to consistently produce high quality partitions for graphs arising in a wide range of application domains. We investigate the effectiveness of many different choices for all three phases: coarsening, partition of the coarsest graph, and refinement. In particular, we present a new coarsening heuristic (called heavy-edge heuristic) for which the size of the partition of the coarse graph is within a small factor of the size of the final partition obtained after multilevel refinement. We also present a much faster variation of the Kernighan--Lin (KL) algorithm for refining during uncoarsening. We test our scheme on a large number of graphs arising in various domains including finite element methods, linear programming, VLSI, and transportation. Our experiments show that our scheme produces partitions that are consistently better than those produced by spectral partitioning schemes in substantially smaller time. Also, when our scheme is used to compute fill-reducing orderings for sparse matrices, it produces orderings that have substantially smaller fill than the widely used multiple minimum degree algorithm.

/pdf/a-fast-and-high-quality-multilevel-scheme-for-partitioning-2i240yved4.pdf

A Fast and High Quality Multilevel Scheme for Partitioning Irregular Graphs

Reliability at massive scale is one of the biggest challenges we face at Amazon.com, one of the largest e-commerce operations in the world; even the slightest outage has significant financial consequences and impacts customer trust. The Amazon.com platform, which provides services for many web sites worldwide, is implemented on top of an infrastructure of tens of thousands of servers and network components located in many datacenters around the world. At this scale, small and large components fail continuously and the way persistent state is managed in the face of these failures drives the reliability and scalability of the software systems.This paper presents the design and implementation of Dynamo, a highly available key-value storage system that some of Amazon's core services use to provide an "always-on" experience. To achieve this level of availability, Dynamo sacrifices consistency under certain failure scenarios. It makes extensive use of object versioning and application-assisted conflict resolution in a manner that provides a novel interface for developers to use.

/pdf/dynamo-amazon-s-highly-available-key-value-store-2ozfk43ou4.pdf

Dynamo: amazon's highly available key-value store

Computational geometry is concerned with the design and analysis of algorithms for geometrical problems. In addition, other more practically oriented, areas of computer science— such as computer graphics, computer-aided design, robotics, pattern recognition, and operations research—give rise to problems that inherently are geometrical. This is one reason computational geometry has attracted enormous research interest in the past decade and is a well-established area today. (For standard sources, we refer to the survey article by Lee and Preparata [19841 and to the textbooks by Preparata and Shames [1985] and Edelsbrunner [1987bl.) Readers familiar with the literature of computational geometry will have noticed, especially in the last few years, an increasing interest in a geometrical construct called the Voronoi diagram. This trend can also be observed in combinatorial geometry and in a considerable number of articles in natural science journals that address the Voronoi diagram under different names specific to the respective area. Given some number of points in the plane, their Voronoi diagram divides the plane according to the nearest-neighbor

Voronoi diagrams—a survey of a fundamental geometric data structure

The success of the von Neumann model of sequential computation is attributable to the fact that it is an efficient bridge between software and hardware: high-level languages can be efficiently compiled on to this model; yet it can be effeciently implemented in hardware. The author argues that an analogous bridge between software and hardware in required for parallel computation if that is to become as widely used. This article introduces the bulk-synchronous parallel (BSP) model as a candidate for this role, and gives results quantifying its efficiency both in implementing high-level language features and algorithms, as well as in being implemented in hardware.

A bridging model for parallel computation

Universal networks offer the advantage that they can execute programs written for simpler architectures without significant run-time overhead. In this paper we investigate simulations of tree machines; the fact that divide-and-conquer algorithms are programmed naturally on trees motivates our investigation. Among various proposals for parallel computing the boolean hypercube has emerged as a particularly versatile network. It is well known that programs for multidimensional grid machines, for example, can be executed on a hypercube with no communications overhead by embedding the grid as a subgraph of the hypercube. Our first result is that a program for any tree machine can be executed on the hypercube with constant overhead. More precisely, every cycle of a synchronous binary tree can be simulated in O(1) cycles on a hypercube, independent of the shape of the tree. The algorithm to embed the tree within the hypercube runs in polynomial time. We also give efficient simulations of arbitrary binary trees on the complete binary tree, the FFT and shuffle-exchange networks. It is natural to ask if any sparse network can simulate every binary tree efficiently. Somewhat surprisingly, we construct a universal bounded-degree network on N nodes for which every N node binary tree is a spanning tree. In other words, every binary tree can be simulated on our universal network with no overhead. This improves previous bounds on the sizes of universal graphs for trees.

Optimal simulations of tree machines

When it comes to achieving performance, reliability, and scalability for commercial-grade Web applications, where is the biggest bottleneck? In many cases today, we see that the limiting bottleneck is the middle mile, or the time data spends traveling back and forth across the Internet, between origin server and end user.

https://dl.acm.org/doi/pdf/10.1145/1466443.1466449

Improving Performance on the Internet: Given the Internet’s bottlenecks, how can we build fast, scalable content-delivery systems?

We describe a polynomial time algorithm that, for every input graph, either outputs the minimum bisection of the graph or halts without output. More importantly, we show that the algorithm chooses the former course with high probability for many natural classes of graphs. In particular, for every fixed d⩾3, all suffciently large n and all b = o(n1-(1/[(d+1)/2]), the algorithm finds the minimum bisection for almost all d-regular labelled simple graphs with 2n nodes and bisection width b.

Graph Bisection Algorithins With Good Average Case Behavior

In this paper, we describe a very simple (1+/spl epsi/)-approximation algorithm for the multicommodity flow problem. The algorithm runs in time that is polynomial in N (the number of nodes in the network) and /spl epsiv//sup -1/ (the closeness of the approximation to optimal). The algorithm is remarkable in that it is much simpler than all known polynomial time flow algorithms (including algorithms for the special case of one-commodity flow). In particular, the algorithm does not rely on augmenting paths, shortest paths, min-cost paths, or similar techniques to push flow through a network. In fact, no explicit attempt is ever made to push flow towards a sink during the algorithm. Because the algorithm is so simple, it can be applied to a variety of problems for which centralized decision making and flow planning is not possible. For example, the algorithm can be easily implemented with local control in a distributed network and it can be made tolerant to link failures. In addition, the algorithm appears to perform well in practice. Initial experiments using the DIMACS generator of test problems indicate that the algorithm performs as well as or better than previously known algorithms, at least for certain test problems. >

/pdf/a-simple-local-control-approximation-algorithm-for-39smqmg9ly.pdf

A simple local-control approximation algorithm for multicommodity flow

(c+d) steps using constant-size queues, where c is the congestion of the paths in the network, and d is the length of the longest path. The proof, however, used the Lovasz Local Lemma and was not constructive. In this paper, we show how to find such a schedule in time, with probability , for any positive constant β, where is the sum of the lengths of the paths taken by the packets in the network, and m is the number of edges used by some packet in the network. We also show how to parallelize the algorithm so that it runs in NC. The method that we use to construct the schedules is based on the algorithmic form of the Lovasz Local Lemma discovered by Beck.

/pdf/fast-algorithms-for-finding-o-congestion-dilation-packet-21oidu8cxw.pdf

Tom Leighton

Papers

Optimal simulations of tree machines

Improving Performance on the Internet: Given the Internet’s bottlenecks, how can we build fast, scalable content-delivery systems?

Graph Bisection Algorithins With Good Average Case Behavior

A simple local-control approximation algorithm for multicommodity flow

Fast Algorithms for Finding O(Congestion + Dilation) Packet Routing Schedules