This is an article I posted over at Lockergnome.
I work at a military headquarters base as a level 2 technician. My responsibilities are broad and diverse, but one of the aspects of my job involves deploying workstations to the users. We use a dual-partition system where the Windows 2000 system files and user profiles are kept on the C: partition and D: is open for data. Approximately 70% of the hard drive is taken up by the D: drive. It’s been that way for so long that no one knows why. Ironically, they did the same thing at my old job, but there, the question was never asked?
On the surface, the idea of a larger empty partition for data seems like a good idea. Data is kept away from Windows so, in the event of a Windows corruption, Windows could be reinstalled. Let’s be realistic though. If Windows is corrupt, I’m going to take 20 minutes and re-image the machine, rather than sit there and try to reinstall Windows through a typical 50 minute installation plus an additional 30 to 40 minutes to configure Windows to a standard configuration. It’s not efficient and opens the network up for problems. That’s why we use an imaging process!!
In reality, however, going dual-partition makes an end-run around the security features inherent in a multi-user environment. In Windows 2000, user data is kept in a profile folder. This folder is accessible only to the user, the local administrators and the domain administrators. If another user signs on to the machine at a later date, he cannot access the data owned by the original user. This is the way things should be.
When a second partition is in place for data storage all data within that drive is accessible by everyone, by default. Obviously, this is not a good thing. In our situation, classified information is not kept on unclassified workstations, but other data is often passed between users legitimately and saved to hard drives. By encouraging users to utilize a second partition accessible by all, instead of a built in profile folder, enterprise managers are encouraging users to access possibly sensitive data. This should be a concern.
There’s history behind the dual-partition method, however. In the context of this history, the method makes sense. But in the context of today’s technology, it doesn’t. Back in the days of Windows NT4 and 8 gigabyte hard drives, system partitions were limited to a maximum of 4 gigs. In those days, if the drive in question was 10gb, the only option was to install a 4gb C: partition and fill the rest with D:. This was due to the NT setup engine formatting a new NTFS partition first as FAT16 which was limited to 4gb, and then converting to NTFS. However, with Windows 2000, this problem does not exist and NTFS partitions can be easily created on the fly. This eliminates the need for two partitions.
Unfortunately, as time goes on and a system is in place longer, the more difficult it is to change the rules in place. Fortunately, our Lieutenant has approved the conversion back to a single partition that preserves the data integrity inherent in a user-specific profile folder. I’ll be implementing these changes gradually, but they will definitely be implemented!
(As a side note, creating and maintaining a small partition, can be useful for speeding up virtual memory and cache and can be used for such a purpose.)
Stumble it!
About the Author: Aaron Brazell is the lead editor of Technosailor.com and a social media expert. His passion is to see companies and individuals use the internet and web technologies wisely and effectively to promote their brands and companies. He served as Director of Technology at b5media from 2005-2008 and is currently an independent consultant.
That makes sense when it comes to multi-user machines, but what about single-user home machines? Is it still a good idea in this case?
On a single user machine, this is not a requirement. It’s not a requirement in enterprise, but I think it’s better. I agree with this emailer who responded to the Lockergnome post, but I only agree in a non-multiuser environment:
No, the reason that dual-partition images are old school is that two is far too few! ;-)
Seriously: I’ve found that Windows machines perform FAR better when the drive is partitioned in smaller pieces. One prominent principle is that you need to keep free space near to where it will be needed. For this reason, you’ll get better performance by putting the user’s profile and major application directories on separate partitions. Partitions can still be linked (via Win2k’s path mounting) to the same place in the file tree, and can be (to neophytes) indistinguishable from directories on the C: partition.
Similary, swap performance can be enhanced by giving swap its own partition near the start of the drive. (Rule of thumb: swap should be near or in the most-used partition on the least-used hard drive.)
E-mail and other major applications will each benefit from having separate partitions, which are now easier to utilize because they can be mounted and made to look like regular directories. I would consider moving the entire “Program Files” tree to its own partition.
And you get double points if you can do this across two or more drives.
The side benefit is that a scrambled root partition should require re-imaging only the root partition, leaving all the user’s profile and other data intact in the separate partitions.
Furthermore, this will reduce the need for defragmentation, thereby giving you an additional day-to-day performance boost.
Try it on your own system. Find the 3 largest directories on your drive (e-mail? user profile? ) and segregate them into separate partitions. (FullDisk http://www.worldlynx.net/pgerhart/_fuldsk.html can help with the analysis.) It will make a big difference in your daily computing experience.