Ioan Raicu

Illinois Institute of Technology

Argonne National Laboratory

CS595: Data-Intensive Computing

Quarter: Fall 2011

Lecture Time: Monday/Wednesday, 11:25AM - 12:40PM

Lecture Location: Stuart Building 106

Professor: Dr. Ioan Raicu (iraicu@cs.iit.edu)

Office Hours Time: Wednesday, 12:45PM - 1:45PM

Office Hours Location: Stuart Building 237D

Teaching Assistant: Siyuan Ma (sma9@hawk.iit.edu)

Office Hours Time: Monday, 12:45PM - 1:45PM

Office Hours Location: Stuart Building 003

This course is a tour through various research topics in distributed data-intensive computing, covering topics in cluster computing, grid computing, supercomputing, and cloud computing. We will explore solutions and learn design principles for building large network-based computational systems to support data intensive computing. This course is geared for junior/senior level undergraduates and graduate students in computer science. 

Lecture topics:

Date Lecture Topic Reading (To be completed by posted date) Assignments
08-22-2011 Syllabus (Slides, PDF)   Reading Write-up Instructions
08-24-2011 Syllabus cont.    
08-29-2011 Introduction to Distributed Systems (Slides)

Reading #1

  • Distributed Computing (HTML)
 
08-31-2011 Introduction to Distributed Systems (cont)    
09-05-2011 Labor Day -- NO CLASS    
09-07-2011 Grid Computing and Cloud Computing (Slides)

Reading #2

  • Cloud Computing and Grid Computing 360-Degree Compared (PDF)
  • Above the clouds: A Berkeley view of cloud computing (PDF)
  • The Anatomy of the Grid (PDF)
  
09-12-2011 Grid Computing and Cloud Computing (cont)

Reading (Optional)

  • What's next in high-performance computing? (PDF)
  • High-Performance Computing (HTML)
  
09-14-2011 Many-Task Computing (Slides)

Reading #3

  • Many-Task Computing for Grids and Supercomputers (PDF)
  • Scalable Parallel Scripting for Scientific Computing (PDF)
  
09-19-2011 Data Intensive Computing Overview (Slides)

Reading #4

  • Foreword (PDF)
  • Jim Gray on eScience (PDF)
 
09-21-2011 Projects Brainstorming (Slides)   Project Proposal (PDF)
09-26-2011 Parallel Programming Systems and Models (Slides)    
09-28-2011 Parallel File Systems (Slides)

Reading #5

  • GPFS: A Shared-Disk File System for Large Computing Clusters (PDF)
  • PVFS: A Parallel File System for Linux Clusters (PDF)
 
10-03-2011 Parallel File Systems

Reading #6

  • Lustre: Building a File System for 1,000-node Clusters (PDF)
  • I/O Performance Challenges at Leadership Scale (PDF)
Project Proposal Due
10-05-2011 Parallel File Systems (Slides)

 

  
10-10-2011 Fall Break - NO CLASS    
10-12-2011 Parallel File Systems

Reading #7

  • Scalable Performance of the Panasas Parallel File System (PDF)
  
10-17-2011 Distributed File Systems (Slides)

Reading #8

  • The Google File System (PDF)
  • On the Duality of Data-intensive File System Design: Reconciling HDFS and PVFS (PDF)
  
10-19-2011 Distributed File Systems

Reading #9

  • Ceph: A Scalable, High-Performance Distributed File System (PDF)
  • Ceph as a scalable alternative to the Hadoop Distributed File System (PDF)
 
10-24-2011 Data Reliability Techniques (Slides)
MapReduce (Slides)

Reading #10

  • A case for redundant arrays of inexpensive disks (RAID) (PDF)
  • A performance evaluation and examination of open-source erasure coding libraries for storage (PDF)

Reading #11

  • MapReduce: Simplified Data Processing on Large Clusters (PDF)
 
10-26-2011 40th Anniversary Celebration -- NO CLASS   Project Mid-term Report (PDF)
10-31-2011 MapReduce

Reading #12

  • A comparison of approaches to large-scale data analysis (PDF)
  • MapReduce: A Flexible Data Processing Tool (PDF)
  • MapReduce and Parallel DBMSs: Friends or Foes? (PDF)

Optional:

  • MapReduce Online (PDF)
  • Building a High-Level Dataflow System on top of Map-Reduce: The Pig Experience (PDF)
 
11-02-2011 Workflow Systems (Slides)

Reading #13

  • Swift: Fast, Reliable, Loosely Coupled Parallel Computation (PDF)
  • Falkon: a Fast and Light-weight tasK executiON framework (PDF)
  • Optional (paper from reading #3): Scalable Parallel Scripting for Scientific Computing (PDF)
Project Mid-term Report Due
11-07-2011 Workflow Systems    
11-09-2011 Distributed Hash Tables (Slides)

Reading #14

  • Dynamo: Amazon’s Highly Available Key-value Store Dynamo (PDF)
  • Kademlia: A Peer-to-peer Information System Based on the XOR Metric (PDF)
 
11-14-2011 Multi/Many Core and MapReduce  -  Small Factories and Big Customers (Slides)

Guest Lecture by Siyuan Ma

Reading #15

  • Improving MapReduce Performance in Heterogeneous Environments (PDF)
  • Server-Side I/O Coordination for Parallel File Systems (PDF)
  • Supporting MapReduce on large-scale asymmetric multi-core clusters (PDF)

Optional

  • IOrchestrator: Improving the Performance of Multi-node I/O Systems via Inter-Server Coordination (PDF)
  • Performance-Driven Task Co-Scheduling for MapReduce Environments (PDF)
  • Reining in the outliers in map-reduce clusters using Mantri (PDF)
  
11-16-2011 Many-core Computing (Slides)

Reading #16

  • Amdahl's law in the multicore era (PDF)
  • Reevaluating Amdahl's Law in the Multicore Era (PDF)
  
11-21-2011 Challenges and Opportunities in Large-Scale Storage Systems (Slides)

Reading #17

  • Making a Case for Distributed File Systems at Exascale (PDF)
  
11-23-2011 Thanksgiving Break -- NO CLASS   Final Project Writeup
11-28-2011
11:25AM-12:45PM		 CLASS CANCELLED    
11-30-2011
11:25AM-12:45PM 		Final Presentations

Final Presentations

  • A Comparison of Distributed Hash Tables, Antonio Perez de Tajada, Rui Han
  • A Distributed Hash Table Implementaion of HDFS & Comparing with Other File Systems, Srinivas Donekal, Venkat Krishna Patibanda, Srikanth Ravindran
  • Performance Evaluation of NFS, PVFS, and HDFS, Yunhui Li, Xu Zhen
 		Feedback
12-02-2011
11:25AM-4:00PM 		Final Presentations

Final Presentations

  • Automatic Paralleliszation of Sequential Code, Prashastha Purushothama, Thierry Silbermann, Vinaykumar Shivakumar
  • Distributed Storage: Rebuilding, Corentin Debains, Pedro Manuel Alvarez-Tabio Togores, Firat Karakusoglu
  • Epidemic Failure Detection in HDFS, Miao Li, Xiao Han, Qi Zhao
  • Implementaion of Centralized-indices for Cloud Database HBASE/Hadoop, Fitsume Dagnew, Jing Guo
  • Distributed Hash Tables and Distributed Filesystems, Wesley Leggette, Patrick O'Shea
    •
  • Performance Maximization of Application-Specific Scheduling for Homogenous Manycore Platform, Ke Yue
  • The Design and Implementation of File System Search Engine Based on the Licene API and HDFS, Dongxun Yin
  • Efficient Computes with General-Purpose Computation on Graphics Processing Units and Many Integrated Core Architecture Technology, Scott Krieder
  • Analysis of Many Task Computing in High End Computing Systems, Prateek Patil
  • Centralized and Distributed Job Scheduling System Simulation at Exascale, Ke Wang
  • Reliability Simulation for Distributed File System at Exascale, Da Zhang
  • An Implementation of Data Provenance in Exascale Distributed Environment, Zhangjie Ma
  • Basic Virtual Network File System: a primitive step for a Low Cost Cloud Storage System, Iman Sandooghi
    •

Final Presentations -- India Section

  • A Comparison of Key-Value based Data Stores, Rohith Kaushik Ananth, Madhumitha Baskaran
  • A Study of MapReduce, Gautam Prasad
 Feedback
12-05-2011 NO CLASS   Project final report due at 11:59PM

Next Semester Spring 2012

In the Spring 2012 semester, I will be teaching CS 553, Cloud Computing.