Download Implementation of Log Based File Systems on Open Source

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Distributed operating system wikipedia , lookup

MTS system architecture wikipedia , lookup

Security-focused operating system wikipedia , lookup

Library (computing) wikipedia , lookup

Plan 9 from Bell Labs wikipedia , lookup

Windows NT startup process wikipedia , lookup

RSTS/E wikipedia , lookup

DNIX wikipedia , lookup

Commodore DOS wikipedia , lookup

OS 2200 wikipedia , lookup

Batch file wikipedia , lookup

Spring (operating system) wikipedia , lookup

Burroughs MCP wikipedia , lookup

Computer file wikipedia , lookup

File locking wikipedia , lookup

VS/9 wikipedia , lookup

CP/M wikipedia , lookup

Unix security wikipedia , lookup

Transcript
Implementation of Log Based File Systems on Open Source Operating Systems
Thesis Proposal by Phil Allen
Overview:
As computing technology advances, specialization occurs. From multi-million dollar
large-scale Cray supercomputers to the booming market of Palm Pilots, each type of computer
has its own set of advantages and disadvantages. Likewise, computing components, both physical
components and software components, have similar sets of advantages and disadvantages
One area of research is the tradeoffs associated with particular types of file systems. The
file system is a software component that directly works with the hardware of the computer; it
determines how and where files are stored on a physical medium such as a hard drive. In the
realm of Windows, there are several file systems available, namely FAT16, FAT32, and NTFS.
In the realm of Open Source Operating Systems, the Fast File System (FFS) is the dominant
operating system, though many others exist due to its open source code nature. The most common
Open Source Operating System is Linux; originally developed by Linus Torvalds, it can be
customized by the user to an amazing degree. One particular brand of Linux is Mandrake, known
for an easy, graphical installation.
A Log File System (LFS) is based on the idea of keeping a log (journal) of all hard drive
reads and writes, preserving old information for a relatively lengthy amount of time. The
advantages of an LFS are 1) memory recovery is extremely easy; by keeping a journal, in the case
of power failure, all but the last few writes to the hard disk are preserved; 2) it is optimized for
writing data; once the position for writing is found, the entire file is written sequentially instead
of fragments being written across the disk as other Operating Systems do; 3) it is optimized for
reading; since each file is written sequentially, once it finds the position to read from, it takes the
absolute minimum, amount of time to read the file.
The main disadvantage of an LFS is the internal component known as a cleaner; it performs much
the same role as a disk defragmenter in Windows. Once the disk has a certain amount of ‘clean’
data, 'dirty' data, and free space, it is necessary to reorganize the 'clean' data so that the
advantages of an LFS are preserved.
Statement of Problem:
The goal of my thesis will be implement a Log File System on Red Hat Linux to a degree
that it is able to perform useful work; doing so will have several benefits. Beyond the general goal
of further educating myself, companies are always looking for reliable ways to store data.
Workstations such as Sun SPARC stations and IBM RS6000 stations have higher hardware
reliability than PCs and that is a contributing reason to the prevalent use of workstations in
scientific and industrial computing. With an LFS, hardware reliability issues are reduced, and as
PCs tend to cost less than workstations, a free LFS on PCs would help enable companies to save
money and/or increase output.
Timetable:

April-September 2001: Reading into research on Log File Systems

April 2001: Learn basics of file systems

May 2001: Implement basic file system as a large file to crystallize basic understanding

June 2001: Convert previously implemented file system to behave as LFS without cleaner
(and error when full)

July 2001: Add basic cleaner to previously implemented file system

August 2001: Begin gaining full understanding of how to create LFS partition type.

September-October 2001: Implement LFS as a Virtual File System (VFS) in RedHat
Linux (without cleaner or with basic cleaner)

October-December 2001: Write thesis while implementation work continues.

November-December 2001: Improve cleaner; fine tune LFS for performance.