CSE 710

Parallel and Distributed File Systems

Spring 2014

 

 

Instructor:

 

Prof. Tevfik Kosar

Office: 338J Davis Hall

Phone: 645-2323

Email: tkosar@buffalo.edu

Office hours: Tuesdays 1:00pm-3:00pm

 

Course Description:

 

As the data requirements of commercial as well as scientific applications continue to increase, the ability to share large amounts of data across widely distributed sites (i.e. data centers, clouds, clusters, supercomputers) becomes more and more important.

 

This seminar will be discussing state-of-the-art research, development, and deployment efforts in parallel and distributed file systems on clustered, grid, and cloud infrastructures. We will be reading and discussing two papers every week in one of the following areas:

 

      File System Design Decisions

      Performance, Scalability, and Consistency issues in File Systems

      Traditional Distributed File Systems

      Parallel Cluster File Systems

      Wide Area Distributed File Systems

      Cloud File Systems

      Commercial vs Open Source File System Solutions

 

Course Location and Time:

 

The seminars will be held Wednesdays 10:00am-12:00pm @ 113A Davis Hall. First day of classes will be on Wednesday, January 29th, 2014.

 

Reading List:

 

The "tentative" reading list for this seminar is available here.

 

Projects:

 

The seminar will include a "term project" on the implementation of the core parallel and distributed file system components, which will be implemented and presented in teams. Some project ideas include:

Project-1: FuseDLS: Design and Implementation of a Fuse-based file system interface to a Cloud-hosted Directory Listing Service
Project-2: MDS: Design and Implementation of a Distributed Metadata Server for Global Name Space in a Wide-area File System
Project-3: SmartFS: Design and Implementation of a Serverless Distributed File System for Smartphones
Project-4: PowerFS: Energy-Aware File System Design
Detailed project descriptions are available here.

Project Milestones are:

Feb 24th: Survey of Related Work
Mar 3rd: System Design Document
Mar 5th: Midterm Presentations
Apr 30th: Final Presentations and Demos
May 12th: Final Reports

 

Grading:

 

This is a research course. There will be team projects and paper presentations, but no exams. Each student will present 1-2 papers from the reading list provided above. Additionally, each student is expected to read all papers, submit questions and comments about the papers, attend classes, and join the discussion of the papers. Grading will be P/F.

 

Useful Links:

 

      How to Read a Paper, by S. Keshav.

      Reviewing a Technical Paper, by M. Ernst

 

Piazza Page:

 

We will also have an online discussion of the papers at the course piazza page. The details about this will be posted here soon.

 

 

Seminar Schedule:


Date

Week

Papers to be Discussed

Presenter

Jan. 29

1

Introduction: Wide Area Distributed File Systems

Tevfik

Feb. 5

2

[1] The Sun Network File System: Design, Implementation and Experience (NFS)

Long

[2] Scale and Performance in a Distributed File System (AFS)

Boyu

Feb. 12

3

[3] Disconnected Operation in the Coda File System

Abhishek

[4] Serverless Network File Systems (xFS)

Alagappan

Feb. 26

4

[5] PVFS: A Parallel File System for Linux Clusters

Mayank

[6] Lustre: A Scalable, High-Performance File System

Aniket

Mar. 5

5

Midterm Presentations

Teams 1 - 3

Mar. 12

6

Midterm Presentations

Teams 4 - 6

Mar. 19

SPRING BREAK

 

 

Mar. 26

7

[7] GPFS: A Shared-Disk File System for Large Computing Clusters

Aditya

[8] Nache: Design and Implementation of a Caching Proxy for NFSv4

Vinay

Apr. 2

8

[9] Panache: A Parallel File System Cache for Global File Access

Jerry

[10] OceanStore: An Architecture for Global-Scale Persistent Storage

Amit

Apr. 9

9

[14] The Google File System

Anuya

[15] Ceph: A Scalable, High-Performance Distributed File System

Sujay

Apr. 16

10

[16] Distributed Directory Service in the Farsite File System

Ankur

[17] Flexible, Wide-Area Storage for Distributed Systems with WheelFS

Sudarshan

Apr. 23

11

[19] BlueSky: A Cloud-Backed File System for the Enterprise

Ketan

[20] Replication, History, and Grafting in the Ori File System

Ailidani

Apr. 30

12

Final Project Presentations & Demos

Teams 4 - 6

May. 7

13

Final Project Presentations & Demos

Teams 1 - 3