G
Guest
A colleague and myself have recently begun a heated
discussion on the benefits of defragging Win XP. He is
using the Inside Microsoft Windows 2000 by David
Solomon. To prove it is worthless. David is formerly a
Consulting Software Engineer at Digital Equipment
Corporation, David worked for over 9 years as a project
leader and developer in the VMS operating system
development group. He developed the Windows NT in OpenVMS
Terms seminar series, which he delivered to over 1000
students worldwide. In addition to organizing and
teaching seminars, David has served as technical chair
for three past Windows NT conferences and has spoken at
WinDev, Windows World, Software Development, four
Microsoft Professional Developers Conferences (PDC), many
Microsoft TechEd's, MSJ Developers Academy, and DECUS. He
was a recipient of the 1993 Microsoft Most Valuable
Professional (MVP) award for Win32 support.
He states in his book that until your Split I/O is above
10 - 20% while running the application of choice,
defragmentation gains you nothing. NTFS creates files
with fragmentation on purpose for better performance.
NTFS expects files to grow, thus putting a 4KB buffer on
each side of the file to allow it to grow, otherwise
fragmentation would be horrid (like FAT12/16/32). It is
a rare event indeed when the disk gets to read or write
contiguously for a long time. It is normal for the heads
to move around as Windows NT reads and writes to
different files in the normal course of its business.
Defragmentation is basically useless.
My argument:
In the Microsoft site under the title of Disk
Defragmenter in Windows 2000: Maintaining Peak
Performance Through Defragmentation, Microsoft virtually
rewrote file system defrag support for Windows XP to
remove the dependency on compressed-file routines and the
Cache Manager. This means that data movement works at
granularity of a single cluster for uncompressed files
and that defragmentation works on NTFS volumes with
cluster sizes larger than 4KB. Also, defragmentation is
now supported on encrypted files.
The other big enhancement is support for online
defragmentation of the MFT and most directory and file
metadata. Finally, there are a number of odd special
cases in the Windows 2000 defragmentation interface that
made writing a defragmenter especially challenging. In
Windows XP, while the defragmentation API interface has
remained unchanged, the way you can use it has improved
enormously, which means better defragmentation that will
result in better system performance. My underline... the
Conclusion states "File fragmentation can negatively
affect operating system speed and performance. To
maintain peak performance when using Windows 2000, the
condition of your disks should be analyzed on a regular
basis-preferably once a week for moderate-to-heavy use;
less frequently for intermittent use-and defragmentation
performed as needed."
Which side has the preponderance of supporting evidence?
Thanks for the long read.
discussion on the benefits of defragging Win XP. He is
using the Inside Microsoft Windows 2000 by David
Solomon. To prove it is worthless. David is formerly a
Consulting Software Engineer at Digital Equipment
Corporation, David worked for over 9 years as a project
leader and developer in the VMS operating system
development group. He developed the Windows NT in OpenVMS
Terms seminar series, which he delivered to over 1000
students worldwide. In addition to organizing and
teaching seminars, David has served as technical chair
for three past Windows NT conferences and has spoken at
WinDev, Windows World, Software Development, four
Microsoft Professional Developers Conferences (PDC), many
Microsoft TechEd's, MSJ Developers Academy, and DECUS. He
was a recipient of the 1993 Microsoft Most Valuable
Professional (MVP) award for Win32 support.
He states in his book that until your Split I/O is above
10 - 20% while running the application of choice,
defragmentation gains you nothing. NTFS creates files
with fragmentation on purpose for better performance.
NTFS expects files to grow, thus putting a 4KB buffer on
each side of the file to allow it to grow, otherwise
fragmentation would be horrid (like FAT12/16/32). It is
a rare event indeed when the disk gets to read or write
contiguously for a long time. It is normal for the heads
to move around as Windows NT reads and writes to
different files in the normal course of its business.
Defragmentation is basically useless.
My argument:
In the Microsoft site under the title of Disk
Defragmenter in Windows 2000: Maintaining Peak
Performance Through Defragmentation, Microsoft virtually
rewrote file system defrag support for Windows XP to
remove the dependency on compressed-file routines and the
Cache Manager. This means that data movement works at
granularity of a single cluster for uncompressed files
and that defragmentation works on NTFS volumes with
cluster sizes larger than 4KB. Also, defragmentation is
now supported on encrypted files.
The other big enhancement is support for online
defragmentation of the MFT and most directory and file
metadata. Finally, there are a number of odd special
cases in the Windows 2000 defragmentation interface that
made writing a defragmenter especially challenging. In
Windows XP, while the defragmentation API interface has
remained unchanged, the way you can use it has improved
enormously, which means better defragmentation that will
result in better system performance. My underline... the
Conclusion states "File fragmentation can negatively
affect operating system speed and performance. To
maintain peak performance when using Windows 2000, the
condition of your disks should be analyzed on a regular
basis-preferably once a week for moderate-to-heavy use;
less frequently for intermittent use-and defragmentation
performed as needed."
Which side has the preponderance of supporting evidence?
Thanks for the long read.