Mizan: A System for Dynamic Load Balancing in Large-scale Graph Processing
|
|
- Blaze Hines
- 5 years ago
- Views:
Transcription
1 /34 Mizan: A System for Dynamic Load Balancing in Large-scale Graph Processing Zuhair Khayyat 1 Karim Awara 1 Amani Alonazi 1 Hani Jamjoom 2 Dan Williams 2 Panos Kalnis 1 1 King Abdullah University of Science and Technology Thuwal, Saudi Arabia 2 IBM Watson Research Center Yorktown Heights, NY
2 2/34 Importance of Graphs Graphs abstract application-specific algorithms into generic problems represented as interactions using vertices and edges Max flow in road network Diameter of WWW Ranking in social networks Simulating protein interactions Algorithms vary in their computational requirements
3 3/34 How to Scale Graph Processing Pregel was introduced by Google as a scalable abstraction for graph processing Overcomes the limitations of processing graphs on MapReduce Based on vertex-centric computation Utilizes bulk synchronous parallel (BSP) System Programming Data Computations Abstraction Exchange map() Key based MapReduce combine() grouping Disk based reduce() compute() Message Pregel combine() passing In memory aggregate()
4 4/34 Pregel s BSP Graph Processing Superstep 1 Superstep 2 Superstep 3 Worker 1 Worker 1 Worker 1 Worker 2 Worker 2 Worker 2 Worker 3 Worker 3 Worker 3 BSP Barrier Balanced computation and communication is fundamental to Pregel s efficiency
5 5/34 Optimizing Graph Processing Existing work focus on optimizing for graph structure (static optimization): Optimize graph partitioning: Simple graph partitioning schemes (e.g., hash or range): Giraph User-defined partitioning function: Pregel Sophisticated partitioning techniques (e.g., min-cuts): GraphLab, PowerGraph, GoldenOrb and Surfer Optimize graph data access: Distributed data stores and graph indexing: GoldenOrb, Hama and Trinity
6 5/34 Optimizing Graph Processing Existing work focus on optimizing for graph structure (static optimization): Optimize graph partitioning: Simple graph partitioning schemes (e.g., hash or range): Giraph User-defined partitioning function: Pregel Sophisticated partitioning techniques (e.g., min-cuts): GraphLab, PowerGraph, GoldenOrb and Surfer Optimize graph data access: Distributed data stores and graph indexing: GoldenOrb, Hama and Trinity What about algorithm behavior? Pregel provides coarse-grained load balancing, is it enough?
7 6/34 Types of Algorithms Algorithms behaves differently, we classify algorithms in two categories depending on their behavior Algorithm Type In/Out Vertices Examples Messages State PageRank, Stationary Predictable Fixed Diameter estimation, WCC Distributed minimum spanning tree, Non-stationary Variable Variable Belief propagation, Graph queries, Ad propagation
8 7/34 Types of Algorithms First Superstep V2 V8 V2 V8 V4 V1 V6 V4 V1 V6 V7 V3 V5 V7 V3 V5 PageRank DMST
9 8/34 Types of Algorithms Superstep k V2 V8 V2 V8 V4 V1 V6 V4 V1 V6 V7 V3 V5 V7 V3 V5 PageRank DMST
10 9/34 Types of Algorithms Superstep k + m V2 V8 V2 V8 V4 V1 V6 V4 V1 V6 V7 V3 V5 V7 V3 V5 PageRank DMST
11 10/34 What Causes Computation Imbalance in Non-stationary Algorithms? Superstep 1 Superstep 2 Superstep 3 Worker 1 Worker 1 Worker 1 Worker 2 Worker 2 Worker 2 Worker 3 Worker 3 Worker 3 BSP Barrier -Vertex response time -Time to receive in messages -Time to send out messages
12 11/34 Why to Optimize for Algorithm Behavior? Difference between stationary and non-stationary algorithms In Messages (Millions) SuperSteps PageRank - Total PageRank - Max/W DMST - Total DMST - Max/W
13 12/34 What is Mizan? BSP-based graph processing framework Uses runtime fine-grained vertex migrations to balance computation and communication Follows the Pregel programming model Open source, written in C++
14 3/34 PageRank s compute() in Mizan void compute(messageiterator * messages, uservertexobject * data, messagemanager * comm) { double currval = data->getvertexvalue().getvalue(); double newval = 0; double c = 0.85; if (data->getcurrentss() > 1) { while (messages->hasnext()) { newval = newval + messages->getnext().getvalue(); } newval = newval * c + (1.0 - c) / ((double) vertextotal); data->setvertexvalue(mdouble(newval)); } else {newval = currval;} if (data->getcurrentss() <= maxsuperstep) { mdouble outval(newval / ((double) data->getoutedgecount())); for (int i = 0; i < data->getoutedgecount(); i++) { comm->sendmessage(data->getoutedgeid(i), outval); } } else {data->votetohalt();} }
15 14/34 Migration Planning Objectives Decentralized Simple Transparent No need to change Pregel s API Does not assume any a priori knowledge to graph structure or algorithm
16 15/34 Mizan s Migration Barrier Mizan performs both planning and migrations after all workers reach the BSP barrier Superstep 1 Superstep 2 Superstep 3 Worker 1 Worker 1 Worker 1 Worker 2 Worker 2 Worker 2 Worker 3 Worker 3 Worker 3 Migration Barrier BSP Barrier Migration Planner
17 16/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance: By comparing the worker s execution time against a normal distribution and flagging outliers
18 16/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance: By comparing the worker s execution time against a normal distribution and flagging outliers Mizan monitors for each vertex: Remote outgoing messages All incoming messages Response time High level summaries are broadcast to each worker Vertex Response Time Remote Outgoing Messages V4 V5 V6 V2 V3 V1 Mizan Worker 1 Local Incoming Messages Mizan Worker 2 Remote Incoming Messages
19 17/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance 2 Select the migration objective: Mizan finds the strongest cause of workload imbalance by comparing statistics for outgoing messages and incoming messages of all workers with the worker s execution time
20 17/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance 2 Select the migration objective: Mizan finds the strongest cause of workload imbalance by comparing statistics for outgoing messages and incoming messages of all workers with the worker s execution time The migration objective is either: Optimize for outgoing messages, or Optimize for incoming messages, or Optimize for response time
21 18/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance 2 Select the migration objective 3 Pair over-utilized workers with under-utilized ones: All workers create and execute the migration plan in parallel without centralized coordination Each worker is paired with one other worker at most W7 W2 W1 W5 W8 W4 W0 W6 W9 W
22 19/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance 2 Select the migration objective 3 Pair over-utilized workers with under-utilized ones 4 Select vertices to migrate: Depending on the migration objective
23 20/34 Mizan s Migration Planning Steps 1 Identify the source of imbalance 2 Select the migration objective 3 Pair over-utilized workers with under-utilized ones 4 Select vertices to migrate 5 Migrate vertices: How to migrate vertices freely across workers while maintaining vertex ownership and fast updates? How to minimize migration costs for large vertices?
24 21/34 Vertex Ownership Mizan uses distributed hash table (DHT) to implement a distributed lookup service: V can execute at any worker V s home worker ID = (hash(id) mod N) Workers ask the home worker of V for its current location The home worker is notified with the new location as V migrates Where is V5 & V8 Compute V1 V3 V2 V5 V8 V4 V2 V5 V6 V9 V7 V1 V3 V9 V8 DHT V1:W1 V4:W2 V7:W3 V2:W1 V5:W2 V8:W3 V3:W1 V6:W2 V9:W3 Worker 1 Worker 2 Worker 3
25 22/34 Migrating Vertices with Large Message Size Introduce delayed migration for very large vertices: At SS t: only ownership of vertex v is moved to worker new Superstep t Superstep t+1 Superstep t+2 Worker_old Worker_old Worker_old ' Worker_new 5 6 Worker_new 5 6 Worker_new
26 22/34 Migrating Vertices with Large Message Size Introduce delayed migration for very large vertices: At SS t: only ownership of vertex v is moved to worker new At SS t + 1: worker new receives vertex 7 messages worker old process vertex 7 Superstep t Superstep t+1 Worker_old Worker_old Worker_old ' Worker_new 5 6 Worker_new 5 6 Worker_new
27 22/34 Migrating Vertices with Large Message Size Introduce delayed migration for very large vertices: At SS t: only ownership of vertex v is moved to worker new At SS t + 1: worker new receives vertex 7 messages worker old process vertex 7 After SS t + 1: worker old moves state of vertex 7 to worker new Superstep t Superstep t+1 Superstep t+2 Worker_old Worker_old Worker_old ' Worker_new 5 6 Worker_new 5 6 Worker_new
28 23/34 Experimental Setup Mizan is implemented on C++ with MPI with three variations: Static Mizan: Emulates Giraph, disables dynamic migration Work stealing (WS): Emulates Pregel s coarse-grained dynamic load balancing Mizan: Activates dynamic migration Local cluster of 21 machines: Mix of i5 and i7 processors with 16GB RAM Each IBM Blue Gene/P supercomputer with 1024 compute nodes: Each is a 4 core PowerPC450 CPU at 850MHz with 4GB RAM
29 24/34 Datasets Experiments on public datasets: Stanford Network Analysis Project (SNAP) The Laboratory for Web Algorithmics (LAW) Kronecker generator name nodes edges kg1 (synthetic) 1,048,576 5,360,368 kg4m68m (synthetic) 4,194,304 68,671,566 web-google 875,713 5,105,039 LiveJournal1 4,847,571 68,993,773 hollywood ,180, ,985,632 arabic ,744, ,999,458
30 25/34 Giraph vs. Static Mizan Run Time (Min) Giraph Static Mizan kg1 web-google kg4m68m LiveJournal1 Run Time (Min) Giraph Static Mizan Vertex count (Millions) Figure: PageRank on social network and random graphs Figure: PageRank on regular random graphs, each has around 17M edge
31 26/34 Effectiveness of Dynamic Vertex Migration PageRank on a social graph (LiveJournal1) The shaded columns: algorithm runtime The unshaded columns: initial partitioning cost 30 Run Time (Min) Static WS Mizan Static WS Mizan Static WS Mizan Hash Range Metis
32 27/34 Effectiveness of Dynamic Vertex Migration Comparing the performance on highly variable messaging pattern algorithms, which are DMST and ad propagation simulation, on a metis partitioned social graph (LiveJournal1) Run Time (Min) Static Work Stealing Mizan 0 DMST Advertisment
33 28/34 Mizan s Overhead with Scaling Speedup 8 Mizan - hollywood Compute Nodes Figure: Speedup on Linux Cluster Speedup Mizan - Arabic Compute Nodes Figure: Speedup on IBM Blue Gene/P supercomputer 1024
34 29/34 Future Work Work with skewed graphs Improving Pregel s fault tolerance
35 30/34 Conclusion Mizan is a Pregel system that uses fine-grained vertex migration to load balance computation and communication across supersteps Mizan is an open source project developed within InfoCloud group in KAUST in collaboration with IBM, programmed in C++ with MPI Mizan scales up to thousands of workers Mizan improves the overall computation cost between 40% up to two order of magnitudes with less than 10% migration overhead Download it at: Try it on EC2 (ami-52ed743b)
36 31/34 Extra: Migration Details and Costs Run Time (Minutes) Superstep Runtime Average Workers Runtime Migration Cost Balance Outgoing Messages Balance Incoming Messages Balance Response Time Supersteps
37 32/34 Extra: Migrated Vertices per Superstep Migrated Vertices (x1000) Migration Cost Total Migrated Vertices Max Vertices Migrated by Single Worker Migration Cost (Minutes) Supersteps
38 33/34 Extra: Mizan s Migration Planning Compute() Migration Barrier No BSP Barrier Superstep Imbalanced? Yes Identify Source of Imbalance Pair with Underutilized Worker Select Vertices and Migrate No Worker Overutilized? Yes
39 34/34 Extra: Architecture of Mizan Mizan Worker Vertex Compute() Migration Planner BSP Processor Communicator - DHT Storage Manager IO HDFS/Local Disks
King Abdullah University of Science and Technology. CS348: Cloud Computing. Large-Scale Graph Processing
King Abdullah University of Science and Technology CS348: Cloud Computing Large-Scale Graph Processing Zuhair Khayyat 10/March/2013 The Importance of Graphs A graph is a mathematical structure that represents
More informationLarge-Scale Graph Processing 1: Pregel & Apache Hama Shiow-yang Wu ( 吳秀陽 ) CSIE, NDHU, Taiwan, ROC
Large-Scale Graph Processing 1: Pregel & Apache Hama Shiow-yang Wu ( 吳秀陽 ) CSIE, NDHU, Taiwan, ROC Lecture material is mostly home-grown, partly taken with permission and courtesy from Professor Shih-Wei
More informationOne Trillion Edges. Graph processing at Facebook scale
One Trillion Edges Graph processing at Facebook scale Introduction Platform improvements Compute model extensions Experimental results Operational experience How Facebook improved Apache Giraph Facebook's
More informationmodern database systems lecture 10 : large-scale graph processing
modern database systems lecture 1 : large-scale graph processing Aristides Gionis spring 18 timeline today : homework is due march 6 : homework out april 5, 9-1 : final exam april : homework due graphs
More informationPregel. Ali Shah
Pregel Ali Shah s9alshah@stud.uni-saarland.de 2 Outline Introduction Model of Computation Fundamentals of Pregel Program Implementation Applications Experiments Issues with Pregel 3 Outline Costs of Computation
More informationPregel: A System for Large- Scale Graph Processing. Written by G. Malewicz et al. at SIGMOD 2010 Presented by Chris Bunch Tuesday, October 12, 2010
Pregel: A System for Large- Scale Graph Processing Written by G. Malewicz et al. at SIGMOD 2010 Presented by Chris Bunch Tuesday, October 12, 2010 1 Graphs are hard Poor locality of memory access Very
More informationDistributed Systems. 21. Graph Computing Frameworks. Paul Krzyzanowski. Rutgers University. Fall 2016
Distributed Systems 21. Graph Computing Frameworks Paul Krzyzanowski Rutgers University Fall 2016 November 21, 2016 2014-2016 Paul Krzyzanowski 1 Can we make MapReduce easier? November 21, 2016 2014-2016
More informationParallel HITS Algorithm Implemented Using HADOOP GIRAPH Framework to resolve Big Data Problem
I J C T A, 9(41) 2016, pp. 1235-1239 International Science Press Parallel HITS Algorithm Implemented Using HADOOP GIRAPH Framework to resolve Big Data Problem Hema Dubey *, Nilay Khare *, Alind Khare **
More informationGPS: A Graph Processing System
GPS: A Graph Processing System Semih Salihoglu and Jennifer Widom Stanford University {semih,widom}@cs.stanford.edu Abstract GPS (for Graph Processing System) is a complete open-source system we developed
More informationECE 7650 Scalable and Secure Internet Services and Architecture ---- A Systems Perspective
ECE 60 Scalable and Secure Internet Services and Architecture ---- A Systems Perspective Part II: Data Center Software Architecture: Topic 3: Programming Models Pregel: A System for Large-Scale Graph Processing
More informationGraph Processing. Connor Gramazio Spiros Boosalis
Graph Processing Connor Gramazio Spiros Boosalis Pregel why not MapReduce? semantics: awkward to write graph algorithms efficiency: mapreduces serializes state (e.g. all nodes and edges) while pregel keeps
More informationGiraph Unchained: Barrierless Asynchronous Parallel Execution in Pregel-like Graph Processing Systems
Giraph Unchained: Barrierless Asynchronous Parallel Execution in Pregel-like Graph Processing Systems University of Waterloo Technical Report CS-215-4 ABSTRACT Minyang Han David R. Cheriton School of Computer
More informationMizan-RMA: Accelerating Mizan Graph Processing Framework with MPI RMA*
216 IEEE 23rd International Conference on High Performance Computing Mizan-RMA: Accelerating Mizan Graph Processing Framework with MPI RMA* Mingzhe Li, Xiaoyi Lu, Khaled Hamidouche, Jie Zhang, and Dhabaleswar
More information[CoolName++]: A Graph Processing Framework for Charm++
[CoolName++]: A Graph Processing Framework for Charm++ Hassan Eslami, Erin Molloy, August Shi, Prakalp Srivastava Laxmikant V. Kale Charm++ Workshop University of Illinois at Urbana-Champaign {eslami2,emolloy2,awshi2,psrivas2,kale}@illinois.edu
More informationFrom Think Like a Vertex to Think Like a Graph. Yuanyuan Tian, Andrey Balmin, Severin Andreas Corsten, Shirish Tatikonda, John McPherson
From Think Like a Vertex to Think Like a Graph Yuanyuan Tian, Andrey Balmin, Severin Andreas Corsten, Shirish Tatikonda, John McPherson Large Scale Graph Processing Graph data is everywhere and growing
More informationAuthors: Malewicz, G., Austern, M. H., Bik, A. J., Dehnert, J. C., Horn, L., Leiser, N., Czjkowski, G.
Authors: Malewicz, G., Austern, M. H., Bik, A. J., Dehnert, J. C., Horn, L., Leiser, N., Czjkowski, G. Speaker: Chong Li Department: Applied Health Science Program: Master of Health Informatics 1 Term
More informationBig Graph Processing. Fenggang Wu Nov. 6, 2016
Big Graph Processing Fenggang Wu Nov. 6, 2016 Agenda Project Publication Organization Pregel SIGMOD 10 Google PowerGraph OSDI 12 CMU GraphX OSDI 14 UC Berkeley AMPLab PowerLyra EuroSys 15 Shanghai Jiao
More informationApache Giraph: Facebook-scale graph processing infrastructure. 3/31/2014 Avery Ching, Facebook GDM
Apache Giraph: Facebook-scale graph processing infrastructure 3/31/2014 Avery Ching, Facebook GDM Motivation Apache Giraph Inspired by Google s Pregel but runs on Hadoop Think like a vertex Maximum value
More informationDistributed Graph Algorithms
Distributed Graph Algorithms Alessio Guerrieri University of Trento, Italy 2016/04/26 This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. Contents 1 Introduction
More informationLarge Scale Graph Processing Pregel, GraphLab and GraphX
Large Scale Graph Processing Pregel, GraphLab and GraphX Amir H. Payberah amir@sics.se KTH Royal Institute of Technology Amir H. Payberah (KTH) Large Scale Graph Processing 2016/10/03 1 / 76 Amir H. Payberah
More informationCS 347 Parallel and Distributed Data Processing
CS 347 Parallel and Distributed Data Processing Spring 2016 Notes 14: Distributed Graph Processing Motivation Many applications require graph processing E.g., PageRank Some graph data sets are very large
More informationGiraph Unchained: Barrierless Asynchronous Parallel Execution in Pregel-like Graph Processing Systems
Giraph Unchained: Barrierless Asynchronous Parallel Execution in Pregel-like Graph Processing Systems ABSTRACT Minyang Han David R. Cheriton School of Computer Science University of Waterloo m25han@uwaterloo.ca
More informationCS 347 Parallel and Distributed Data Processing
CS 347 Parallel and Distributed Data Processing Spring 2016 Notes 14: Distributed Graph Processing Motivation Many applications require graph processing E.g., PageRank Some graph data sets are very large
More informationGiraph: Large-scale graph processing infrastructure on Hadoop. Qu Zhi
Giraph: Large-scale graph processing infrastructure on Hadoop Qu Zhi Why scalable graph processing? Web and social graphs are at immense scale and continuing to grow In 2008, Google estimated the number
More informationTI2736-B Big Data Processing. Claudia Hauff
TI2736-B Big Data Processing Claudia Hauff ti2736b-ewi@tudelft.nl Intro Streams Streams Map Reduce HDFS Pig Ctd. Graphs Pig Design Patterns Hadoop Ctd. Giraph Zoo Keeper Spark Spark Ctd. Learning objectives
More informationAutomatic Scaling Iterative Computations. Aug. 7 th, 2012
Automatic Scaling Iterative Computations Guozhang Wang Cornell University Aug. 7 th, 2012 1 What are Non-Iterative Computations? Non-iterative computation flow Directed Acyclic Examples Batch style analytics
More informationPREGEL: A SYSTEM FOR LARGE- SCALE GRAPH PROCESSING
PREGEL: A SYSTEM FOR LARGE- SCALE GRAPH PROCESSING G. Malewicz, M. Austern, A. Bik, J. Dehnert, I. Horn, N. Leiser, G. Czajkowski Google, Inc. SIGMOD 2010 Presented by Ke Hong (some figures borrowed from
More informationHigh Performance Data Analytics: Experiences Porting the Apache Hama Graph Analytics Framework to an HPC InfiniBand Connected Cluster
High Performance Data Analytics: Experiences Porting the Apache Hama Graph Analytics Framework to an HPC InfiniBand Connected Cluster Summary Open source analytic frameworks, such as those in the Apache
More informationCS /21/2016. Paul Krzyzanowski 1. Can we make MapReduce easier? Distributed Systems. Apache Pig. Apache Pig. Pig: Loading Data.
Distributed Systems 1. Graph Computing Frameworks Can we make MapReduce easier? Paul Krzyzanowski Rutgers University Fall 016 1 Apache Pig Apache Pig Why? Make it easy to use MapReduce via scripting instead
More informationOn Smart Query Routing: For Distributed Graph Querying with Decoupled Storage
On Smart Query Routing: For Distributed Graph Querying with Decoupled Storage Arijit Khan Nanyang Technological University (NTU), Singapore Gustavo Segovia ETH Zurich, Switzerland Donald Kossmann Microsoft
More informationPAGE: A Partition Aware Graph Computation Engine
PAGE: A Graph Computation Engine Yingxia Shao, Junjie Yao, Bin Cui, Lin Ma Department of Computer Science Key Lab of High Confidence Software Technologies (Ministry of Education) Peking University {simon227,
More informationAN EXPERIMENTAL COMPARISON OF PARTITIONING STRATEGIES IN DISTRIBUTED GRAPH PROCESSING SHIV VERMA THESIS
c 2017 Shiv Verma AN EXPERIMENTAL COMPARISON OF PARTITIONING STRATEGIES IN DISTRIBUTED GRAPH PROCESSING BY SHIV VERMA THESIS Submitted in partial fulfillment of the requirements for the degree of Master
More informationTurning NoSQL data into Graph Playing with Apache Giraph and Apache Gora
Turning NoSQL data into Graph Playing with Apache Giraph and Apache Gora Team Renato Marroquín! PhD student: Interested in: Information retrieval. Distributed and scalable data management. Apache Gora:
More informationGraph-Processing Systems. (focusing on GraphChi)
Graph-Processing Systems (focusing on GraphChi) Recall: PageRank in MapReduce (Hadoop) Input: adjacency matrix H D F S (a,[c]) (b,[a]) (c,[a,b]) (c,pr(a) / out (a)), (a,[c]) (a,pr(b) / out (b)), (b,[a])
More informationComputation and Communication Efficient Graph Processing with Distributed Immutable View
Computation and Communication Efficient Graph Processing with Distributed Immutable View Rong Chen, Xin Ding, Peng Wang, Haibo Chen, Binyu Zang, Haibing Guan Shanghai Key Laboratory of Scalable Computing
More informationGraph-Parallel Problems. ML in the Context of Parallel Architectures
Case Study 4: Collaborative Filtering Graph-Parallel Problems Synchronous v. Asynchronous Computation Machine Learning for Big Data CSE547/STAT548, University of Washington Emily Fox February 20 th, 2014
More informationFlashGraph: Processing Billion-Node Graphs on an Array of Commodity SSDs
FlashGraph: Processing Billion-Node Graphs on an Array of Commodity SSDs Da Zheng, Disa Mhembere, Randal Burns Department of Computer Science Johns Hopkins University Carey E. Priebe Department of Applied
More informationCOSC 6339 Big Data Analytics. Graph Algorithms and Apache Giraph
COSC 6339 Big Data Analytics Graph Algorithms and Apache Giraph Parts of this lecture are adapted from UMD Jimmy Lin s slides, which is licensed under a Creative Commons Attribution-Noncommercial-Share
More informationPREGEL: A SYSTEM FOR LARGE-SCALE GRAPH PROCESSING
PREGEL: A SYSTEM FOR LARGE-SCALE GRAPH PROCESSING Grzegorz Malewicz, Matthew Austern, Aart Bik, James Dehnert, Ilan Horn, Naty Leiser, Grzegorz Czajkowski (Google, Inc.) SIGMOD 2010 Presented by : Xiu
More informationBSP, Pregel and the need for Graph Processing
BSP, Pregel and the need for Graph Processing Patrizio Dazzi, HPC Lab ISTI - CNR mail: patrizio.dazzi@isti.cnr.it web: http://hpc.isti.cnr.it/~dazzi/ National Research Council of Italy A need for Graph
More informationResearch challenges in data-intensive computing The Stratosphere Project Apache Flink
Research challenges in data-intensive computing The Stratosphere Project Apache Flink Seif Haridi KTH/SICS haridi@kth.se e2e-clouds.org Presented by: Seif Haridi May 2014 Research Areas Data-intensive
More informationLogGP: A Log-based Dynamic Graph Partitioning Method
LogGP: A Log-based Dynamic Graph Partitioning Method Ning Xu, Lei Chen, Bin Cui Department of Computer Science, Peking University, Beijing, China Hong Kong University of Science and Technology, Hong Kong,
More informationDistributed Graph Storage. Veronika Molnár, UZH
Distributed Graph Storage Veronika Molnár, UZH Overview Graphs and Social Networks Criteria for Graph Processing Systems Current Systems Storage Computation Large scale systems Comparison / Best systems
More informationResilient Distributed Datasets
Resilient Distributed Datasets A Fault- Tolerant Abstraction for In- Memory Cluster Computing Matei Zaharia, Mosharaf Chowdhury, Tathagata Das, Ankur Dave, Justin Ma, Murphy McCauley, Michael Franklin,
More informationWhy do we need graph processing?
Why do we need graph processing? Community detection: suggest followers? Determine what products people will like Count how many people are in different communities (polling?) Graphs are Everywhere Group
More informationHIGH PERFORMANCE DATA ANALYTICS:
www.gdmissionsystems.com/hpc HIGH PERFORMANCE DATA ANALYTICS: Experiences Porting the Apache Hama Graph Analytics Framework to an HPC InfiniBand Connected Cluster 1. Summary Open source analytic frameworks,
More informationPREGEL. A System for Large Scale Graph Processing
PREGEL A System for Large Scale Graph Processing The Problem Large Graphs are often part of computations required in modern systems (Social networks and Web graphs etc.) There are many graph computing
More informationGraph Partitioning for Scalable Distributed Graph Computations
Graph Partitioning for Scalable Distributed Graph Computations Aydın Buluç ABuluc@lbl.gov Kamesh Madduri madduri@cse.psu.edu 10 th DIMACS Implementation Challenge, Graph Partitioning and Graph Clustering
More informationAccelerating Irregular Computations with Hardware Transactional Memory and Active Messages
MACIEJ BESTA, TORSTEN HOEFLER spcl.inf.ethz.ch Accelerating Irregular Computations with Hardware Transactional Memory and Active Messages LARGE-SCALE IRREGULAR GRAPH PROCESSING Becoming more important
More informationNavigating the Maze of Graph Analytics Frameworks using Massive Graph Datasets
Navigating the Maze of Graph Analytics Frameworks using Massive Graph Datasets Nadathur Satish, Narayanan Sundaram, Mostofa Ali Patwary, Jiwon Seo, Jongsoo Park, M. Amber Hassaan, Shubho Sengupta, Zhaoming
More informationGiraphx: Parallel Yet Serializable Large-Scale Graph Processing
Giraphx: Parallel Yet Serializable Large-Scale Graph Processing Serafettin Tasci and Murat Demirbas Computer Science & Engineering Department University at Buffalo, SUNY Abstract. Bulk Synchronous Parallelism
More informationPiccolo. Fast, Distributed Programs with Partitioned Tables. Presenter: Wu, Weiyi Yale University. Saturday, October 15,
Piccolo Fast, Distributed Programs with Partitioned Tables 1 Presenter: Wu, Weiyi Yale University Outline Background Intuition Design Evaluation Future Work 2 Outline Background Intuition Design Evaluation
More informationThe Future of High Performance Computing
The Future of High Performance Computing Randal E. Bryant Carnegie Mellon University http://www.cs.cmu.edu/~bryant Comparing Two Large-Scale Systems Oakridge Titan Google Data Center 2 Monolithic supercomputer
More informationGraphHP: A Hybrid Platform for Iterative Graph Processing
GraphHP: A Hybrid Platform for Iterative Graph Processing Qun Chen, Song Bai, Zhanhuai Li, Zhiying Gou, Bo Suo and Wei Pan Northwestern Polytechnical University Xi an, China {chenbenben, baisong, lizhh,
More informationExtreme-scale Graph Analysis on Blue Waters
Extreme-scale Graph Analysis on Blue Waters 2016 Blue Waters Symposium George M. Slota 1,2, Siva Rajamanickam 1, Kamesh Madduri 2, Karen Devine 1 1 Sandia National Laboratories a 2 The Pennsylvania State
More informationAn Experimental Comparison of Partitioning Strategies in Distributed Graph Processing
An Experimental Comparison of Partitioning Strategies in Distributed Graph Processing ABSTRACT Shiv Verma 1, Luke M. Leslie 1, Yosub Shin, Indranil Gupta 1 1 University of Illinois at Urbana-Champaign,
More informationThe Fusion Distributed File System
Slide 1 / 44 The Fusion Distributed File System Dongfang Zhao February 2015 Slide 2 / 44 Outline Introduction FusionFS System Architecture Metadata Management Data Movement Implementation Details Unique
More informationPREGEL AND GIRAPH. Why Pregel? Processing large graph problems is challenging Options
Data Management in the Cloud PREGEL AND GIRAPH Thanks to Kristin Tufte 1 Why Pregel? Processing large graph problems is challenging Options Custom distributed infrastructure Existing distributed computing
More informationAn Algorithmic Approach to Communication Reduction in Parallel Graph Algorithms
An Algorithmic Approach to Communication Reduction in Parallel Graph Algorithms Harshvardhan, Adam Fidel, Nancy M. Amato, Lawrence Rauchwerger Parasol Laboratory Dept. of Computer Science and Engineering
More informationGraph Data Management
Graph Data Management Analysis and Optimization of Graph Data Frameworks presented by Fynn Leitow Overview 1) Introduction a) Motivation b) Application for big data 2) Choice of algorithms 3) Choice of
More informationMapReduce: A Programming Model for Large-Scale Distributed Computation
CSC 258/458 MapReduce: A Programming Model for Large-Scale Distributed Computation University of Rochester Department of Computer Science Shantonu Hossain April 18, 2011 Outline Motivation MapReduce Overview
More informationPlanar: Parallel Lightweight Architecture-Aware Adaptive Graph Repartitioning
Planar: Parallel Lightweight Architecture-Aware Adaptive Graph Repartitioning Angen Zheng, Alexandros Labrinidis, and Panos K. Chrysanthis University of Pittsburgh 1 Graph Partitioning Applications of
More informationNXgraph: An Efficient Graph Processing System on a Single Machine
NXgraph: An Efficient Graph Processing System on a Single Machine Yuze Chi, Guohao Dai, Yu Wang, Guangyu Sun, Guoliang Li and Huazhong Yang Tsinghua National Laboratory for Information Science and Technology,
More informationExtreme-scale Graph Analysis on Blue Waters
Extreme-scale Graph Analysis on Blue Waters 2016 Blue Waters Symposium George M. Slota 1,2, Siva Rajamanickam 1, Kamesh Madduri 2, Karen Devine 1 1 Sandia National Laboratories a 2 The Pennsylvania State
More informationMapReduce Spark. Some slides are adapted from those of Jeff Dean and Matei Zaharia
MapReduce Spark Some slides are adapted from those of Jeff Dean and Matei Zaharia What have we learnt so far? Distributed storage systems consistency semantics protocols for fault tolerance Paxos, Raft,
More informationPuLP: Scalable Multi-Objective Multi-Constraint Partitioning for Small-World Networks
PuLP: Scalable Multi-Objective Multi-Constraint Partitioning for Small-World Networks George M. Slota 1,2 Kamesh Madduri 2 Sivasankaran Rajamanickam 1 1 Sandia National Laboratories, 2 The Pennsylvania
More informationGraph Processing Frameworks
Graph Processing Frameworks Lecture 24 CSCI 4974/6971 5 Dec 2016 1 / 13 Today s Biz 1. Reminders 2. Review 3. Graph Processing Frameworks 4. 2D Partitioning 2 / 13 Reminders Assignment 6: due date Dec
More informationLatency-Tolerant Software Distributed Shared Memory
Latency-Tolerant Software Distributed Shared Memory Jacob Nelson, Brandon Holt, Brandon Myers, Preston Briggs, Luis Ceze, Simon Kahan, Mark Oskin University of Washington USENIX ATC 2015 July 9, 2015 25
More informationOptimizing CPU Cache Performance for Pregel-Like Graph Computation
Optimizing CPU Cache Performance for Pregel-Like Graph Computation Songjie Niu, Shimin Chen* State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences
More informationExperimental Analysis of Distributed Graph Systems
Experimental Analysis of Distributed Graph ystems Khaled Ammar, M. Tamer Özsu David R. Cheriton chool of Computer cience University of Waterloo, Waterloo, Ontario, Canada {khaled.ammar, tamer.ozsu}@uwaterloo.ca
More informationDeveloping MapReduce Programs
Cloud Computing Developing MapReduce Programs Dell Zhang Birkbeck, University of London 2017/18 MapReduce Algorithm Design MapReduce: Recap Programmers must specify two functions: map (k, v) * Takes
More informationGraph Analytics using Vertica Relational Database
Graph Analytics using ertica Relational Database Alekh Jindal* Samuel Madden Malú Castellanos Meichun Hsu Microsoft MIT ertica ertica * work done while at MIT Motivation for graphs on DB Data anyways in
More informationhyperx: scalable hypergraph processing
hyperx: scalable hypergraph processing Jin Huang November 15, 2015 The University of Melbourne overview Research Outline Scalable Hypergraph Processing Problem and Challenge Idea Solution Implementation
More informationNXgraph: An Efficient Graph Processing System on a Single Machine
NXgraph: An Efficient Graph Processing System on a Single Machine arxiv:151.91v1 [cs.db] 23 Oct 215 Yuze Chi, Guohao Dai, Yu Wang, Guangyu Sun, Guoliang Li and Huazhong Yang Tsinghua National Laboratory
More informationPregel: A System for Large-Scale Graph Proces sing
Pregel: A System for Large-Scale Graph Proces sing Grzegorz Malewicz, Matthew H. Austern, Aart J. C. Bik, James C. Dehnert, Ilan Horn, Naty Leiser, and Grzegorz Czajkwoski Google, Inc. SIGMOD July 20 Taewhi
More informationMapReduce, Hadoop and Spark. Bompotas Agorakis
MapReduce, Hadoop and Spark Bompotas Agorakis Big Data Processing Most of the computations are conceptually straightforward on a single machine but the volume of data is HUGE Need to use many (1.000s)
More informationManaging and Mining Billion Node Graphs. Haixun Wang Microsoft Research Asia
Managing and Mining Billion Node Graphs Haixun Wang Microsoft Research Asia Outline Overview Storage Online query processing Offline graph analytics Advanced applications Is it hard to manage graphs? Good
More informationPuLP. Complex Objective Partitioning of Small-World Networks Using Label Propagation. George M. Slota 1,2 Kamesh Madduri 2 Sivasankaran Rajamanickam 1
PuLP Complex Objective Partitioning of Small-World Networks Using Label Propagation George M. Slota 1,2 Kamesh Madduri 2 Sivasankaran Rajamanickam 1 1 Sandia National Laboratories, 2 The Pennsylvania State
More informationToday s content. Resilient Distributed Datasets(RDDs) Spark and its data model
Today s content Resilient Distributed Datasets(RDDs) ------ Spark and its data model Resilient Distributed Datasets: A Fault- Tolerant Abstraction for In-Memory Cluster Computing -- Spark By Matei Zaharia,
More informationDNA Interaction Network
Social Network Web Network Social Network DNA Interaction Network Follow Network User-Product Network Nonuniform network comm costs Contentiousness of the memory subsystems Nonuniform comp requirement
More informationThe Stratosphere Platform for Big Data Analytics
The Stratosphere Platform for Big Data Analytics Hongyao Ma Franco Solleza April 20, 2015 Stratosphere Stratosphere Stratosphere Big Data Analytics BIG Data Heterogeneous datasets: structured / unstructured
More informationGLADE: A Scalable Framework for Efficient Analytics. Florin Rusu University of California, Merced
GLADE: A Scalable Framework for Efficient Analytics Florin Rusu University of California, Merced Motivation and Objective Large scale data processing Map-Reduce is standard technique Targeted to distributed
More informationShark. Hive on Spark. Cliff Engle, Antonio Lupher, Reynold Xin, Matei Zaharia, Michael Franklin, Ion Stoica, Scott Shenker
Shark Hive on Spark Cliff Engle, Antonio Lupher, Reynold Xin, Matei Zaharia, Michael Franklin, Ion Stoica, Scott Shenker Agenda Intro to Spark Apache Hive Shark Shark s Improvements over Hive Demo Alpha
More informationA Cost-efficient Auto-scaling Algorithm for Large-scale Graph Processing in Cloud Environments with Heterogeneous Resources
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, MANUSCRIPT ID 1 A Cost-efficient Auto-scaling Algorithm for Large-scale Graph Processing in Cloud Environments with Heterogeneous Resources Safiollah Heidari,
More informationGiViP: A Visual Profiler for Distributed Graph Processing Systems
GiViP: A Visual Profiler for Distributed Graph Processing Systems Alessio Arleo, Walter Didimo, Giuseppe Liotta and Fabrizio Montecchiani University of Perugia, Italy The Value of Big Graphs The analysis
More informationLFGraph: Simple and Fast Distributed Graph Analytics
LFGraph: Simple and Fast Distributed Graph Analytics Imranul Hoque VMware, Inc. ihoque@vmware.com Indranil Gupta University of Illinois, Urbana-Champaign indy@illinois.edu Abstract Distributed graph analytics
More informationOn Fast Parallel Detection of Strongly Connected Components (SCC) in Small-World Graphs
On Fast Parallel Detection of Strongly Connected Components (SCC) in Small-World Graphs Sungpack Hong 2, Nicole C. Rodia 1, and Kunle Olukotun 1 1 Pervasive Parallelism Laboratory, Stanford University
More informationThe Bulk Synchronous Parallel Model (PSC 1.2) Lecture 1.2 Bulk Synchronous Parallel Model p.1
The Bulk Synchronous Parallel Model (PSC 1.2) Lecture 1.2 Bulk Synchronous Parallel Model p.1 What is a parallel computer? switch switch switch switch PC PC PC PC PC PC PC PC A parallel computer consists
More informationJure Leskovec Including joint work with Y. Perez, R. Sosič, A. Banarjee, M. Raison, R. Puttagunta, P. Shah
Jure Leskovec (@jure) Including joint work with Y. Perez, R. Sosič, A. Banarjee, M. Raison, R. Puttagunta, P. Shah 2 My research group at Stanford: Mining and modeling large social and information networks
More informationMassive Online Analysis - Storm,Spark
Massive Online Analysis - Storm,Spark presentation by R. Kishore Kumar Research Scholar Department of Computer Science & Engineering Indian Institute of Technology, Kharagpur Kharagpur-721302, India (R
More informationGraph Processing & Bulk Synchronous Parallel Model
Graph Processing & Bulk Synchronous Parallel Model CompSci 590.03 Instructor: Ashwin Machanavajjhala Lecture 14 : 590.02 Spring 13 1 Recap: Graph Algorithms Many graph algorithms need iterafve computafon
More informationParallel Computing: MapReduce Jin, Hai
Parallel Computing: MapReduce Jin, Hai School of Computer Science and Technology Huazhong University of Science and Technology ! MapReduce is a distributed/parallel computing framework introduced by Google
More informationGridGraph: Large-Scale Graph Processing on a Single Machine Using 2-Level Hierarchical Partitioning. Xiaowei ZHU Tsinghua University
GridGraph: Large-Scale Graph Processing on a Single Machine Using -Level Hierarchical Partitioning Xiaowei ZHU Tsinghua University Widely-Used Graph Processing Shared memory Single-node & in-memory Ligra,
More informationLecture 22 : Distributed Systems for ML
10-708: Probabilistic Graphical Models, Spring 2017 Lecture 22 : Distributed Systems for ML Lecturer: Qirong Ho Scribes: Zihang Dai, Fan Yang 1 Introduction Big data has been very popular in recent years.
More informationParallel Processing of Large Graphs. Tomasz Kajdanowicz, Przemyslaw Kazienko, Wojciech Indyk. Wroclaw University of Technology, Poland
Parallel Processing of Large Graphs Tomasz Kajdanowicz, Przemyslaw Kazienko, Wojciech Indyk Wroclaw University of Technology, Poland arxiv:1306.0326v1 [cs.dc] 3 Jun 2013 Abstract More and more large data
More informationCluster Computing Architecture. Intel Labs
Intel Labs Legal Notices INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED
More informationBig Data Infrastructure CS 489/698 Big Data Infrastructure (Winter 2016)
Big Data Infrastructure CS 489/698 Big Data Infrastructure (Winter 2016) Week 2: MapReduce Algorithm Design (2/2) January 14, 2016 Jimmy Lin David R. Cheriton School of Computer Science University of Waterloo
More informationDistributed Systems. 20. Other parallel frameworks. Paul Krzyzanowski. Rutgers University. Fall 2017
Distributed Systems 20. Other parallel frameworks Paul Krzyzanowski Rutgers University Fall 2017 November 20, 2017 2014-2017 Paul Krzyzanowski 1 Can we make MapReduce easier? 2 Apache Pig Why? Make it
More informationIrregular Graph Algorithms on Parallel Processing Systems
Irregular Graph Algorithms on Parallel Processing Systems George M. Slota 1,2 Kamesh Madduri 1 (advisor) Sivasankaran Rajamanickam 2 (Sandia mentor) 1 Penn State University, 2 Sandia National Laboratories
More informationCS November 2017
Distributed Systems 0. Other parallel frameworks Can we make MapReduce easier? Paul Krzyzanowski Rutgers University Fall 017 November 0, 017 014-017 Paul Krzyzanowski 1 Apache Pig Apache Pig Why? Make
More information