Wiki code for Proposal: ZFS

Hide Line numbers
1: == A Study of the Zettabyte File System (ZFS)
2: 
3: ==== Kurchi Subhra Hazra - lead student researcher  Goutam Sanyal - faculty advisor  Tania Hamid - student researcher
4: 
5: === National Institute of Technology, Durgapur, West Bengal, India
6: 
7: [[Second Progress Report>>attach:Project awards.files@ZFSProgressReport1.pdf]] [PDF]
8: 
9: [[First Progress Report>>attach:Project awards.files@ZFSProgressReport0.pdf]] [PDF]
10: 
11: ==== Statement of Purpose:
12: 
13: The research work is aimed at carrying out a thorough study of Zettabyte File System, a new kind of file system developed by Sun Microsystems for Solaris OS, which claims to provide simple administration, transactional semantics, end-to-end data integrity, and immense scalability. We propose verifying its performance and explain the working behind it. As an ambitious proposal, we also intend to suggest some possible improvements in it.
14: 
15: ==== Abstract
16: 
17: ZFS is a new kind of file system that provides simple administration, transactional semantics, end-to-end data integrity, and immense scalability. ZFS is not an incremental improvement to existing technology; it is a fundamentally new approach to data management. Although ZFS has is available on all Solaris 10 OS-supported platforms now, there has not been much of work to explain and verify every detail of it.
18: 
19:  Some work has already been done in this connection though. There exists white papers like “Solaris ZFS vs. Red Hat Enterprise Linux EXT3 White Paper”, “Solaris ZFS and Microsoft Windows Server 2003 NTFS File System Performance” and “SOLARIS™ ZFS AND VERITAS STORAGE FOUNDATION FILE SYSTEM PERFORMANCE White Paper June 2007 to compare the performance of ZFS to other file systems.
20: 
21: ==== Key technology goals of the project
22: 
23: * Carry out a thorough study of ZFS
24: * Compare and contrast it with other file systems
25: * Write a thorough one stop document as a ZFS reference material
26: 
27: ==== Research Methodology
28: 
29: **Module 1: A thorough conceptual study**
30: 
31: **Time: 12 weeks**
32: 
33: Initially, a thorough study of ZFS is to be carried out. We intend to install Solaris 10 or use a Sun Workstation, as grants permit, and study the working of ZFS. We plan to verify the following claimed characteristics of ZFS :
34: 
35: Pooled storage
36: 
37: * Completely eliminates the antique notion of volumes
38: * Does for storage what VM did for memory
39: 
40: Transactional object system
41: 
42: * Always consistent on disk – no fsck, ever
43: * Universal – file, block, iSCSI, swap ...
44: 
45: Provable end-to-end data integrity
46: 
47: * Detects and corrects silent data corruption
48: * Historically considered “too expensive” – no longer true
49: 
50: Simple administration
51: 
52: * Concisely express your intent
53: 
54: We also take a thorough look into the compression standards used in ZFS as well as the encryption techniques used.
55: 
56: **Module 2: A thorough comparative study of ZFS verses other File Systems**
57: 
58: **Time: 6 weeks**
59: 
60: In this phase, a comparative study of ZFS with other file systems ( Linux Ext 3 and NTFS) will be carried out, to bring to light the advantages of ZFS over other file systems. All concepts behind ZFS will be compared and contrasted with that of the other file systems mentioned.
61: 
62: **Module 3: Comprehensive All encompassing Document on ZFS and possible suggestion for improvements in ZFS**
63: 
64: **Time: 4 weeks**
65: 
66:  In this phase, we plan to put together all our study under one document(in pdf format) that will serve as a detailed reference for ZFS. The document will be comprehensive, as well as, well illustrated easy to read.
67: 
68:  Besides, as time permits, based on our research work, we would also like to point out possible suggestions to improve the file system.
69: 
70:  The grant from OpenSolaris community will help us in purchasing any required systems(Sun work station) or softwares (Solaris 10/Linux/Other softwares required for performance evaluation) that will be required to carry out the work in the given 3 phases.
71: 
72: ==== Expected outcome of this research
73: 
74: The proposed research work will be beneficial to the entire Open Solaris community. A simple search shows the lack of knowledge resources on ZFS. We thus intend to come up with a detailed and comprehensive document on ZFS, the concepts behind it and its advantages.
75: 
76: ==== Significant Deliverables:
77: 
78: * A complete reference document on ZFS in pdf format.
79: * We also plan to hold talks and seminars on campus to make them aware of the file system and its advantages.
80: 
81: ==== References
82: 
83: 1. http://www.opensolaris.org/os/community/zfs/
84: 1. http://www.sun.com/2004-0914/feature/
85: 1. http://www.opensolaris.org/os/community/zfs/whatis/
86: 1. White Paper: Solaris ZFS and Red Hat Enterprise Linux ext3 File System Performance
87: 1. http://blogs.sun.com/roch/entry/the_dynamics_of_zfs
88: 1. Introduction to Operating Systems: A Hands-On ApproachUsing the OpenSolaris Project Student Guide
OpenSolaris

Top Menu

Wiki code for Proposal: ZFS

Search

Collectives

Community Group

Project

User Group

Subsites

Project awards Pages
