Hello

I'm not starting a topic FAT32 vs NTFS because NTFS is far better due to security and hotfixing. We know it all.
I'm a systemoperator and I've some serious questions/remarks about FAT32. I don't know exactly if I'm right, otherwise I don't make this topic.

FAT32 is outdated and on modern harddisk you need 32KB units or otherwise the FAT can't ben cached. FAT32 has no security in mind and a multi account system is not considerable with FAT32 so don't use it, but why am I still using it?

My system:
AMD64 3800+ @2880Mhz
2048 MB RAM (2 sticks became corrupted and are RTM)
2 IDE Harddisk (one of 40 and one of 80 GB)(enough for me)
ATi X1950XT graphics card (1680x1050@4xAA is no issue)

1. The security overlay from NTFS is a huge overhead. I'm using a harddisk access monitoring system and on X64 with FAT32 the avarage is 312 IO/s and in Vista with NTFS it whas 1455 IO/s. You understand hopefully that a huge IO load on the harddisk shorten it's life.
2. Both harddisk stays 8 degrees Celsius cooler than under Vista. That's scary. (I checked this out. Now there're just cold)
3. My system is still not infected yet. If FAT32 is still that unsafe, my system should be infected. Why this isn't happening (Not I like it, but what function uses the rights management of NTFS for homeusers?) In corporate and industrial surroundings this discussion is not any concern of. NTFS is the way (UNICOS (Unix) were better in history than NTFS ever could).
4. Lost clusters are being stored in files and on NTFS this data is lost or rolled back-up to the previous one. The lost clusters were clusters being written once the error occurs. It is the most recent data and full recoverable till the last bit written (Why FAT is not recoverable??). Nor NTFS could protect you from these single point of faillures. It simply rolls your action back and the data is between 1 and 16 seconds outdated due to the roll-back. (learn about NTFS at Wikipedia eg). For the homeuser not an issue, but we called it: lazy writing. FAT32 don't support any kind of lazy writing and a writing command is executed immediately (or the harddiskcache traps in, but rather than Windows).
5. Windows X64 behaviour on FAT32 is excellent, sorry, I mean this seriously. It flies like never before.
6. FAT is far more friendly to SSD's than NTFS with a fixed log which using it constanly.
7. exFAT is the successor for USB devices with Vista's SP1 which integrates it. Still FAT.............

The disadvantages why not using it:

1. At servers FAT is not any concern of. It may not be existing on any server in the world. point.
2. If you share your system (family) and each one have inportant data, stickj with NTFS with no doubt.
3. Permissions and Rights are most needed by homeusers, it should protect them for their unknowness. But, not every user is a beginner. My skilles aren't perfect, but good enough to keep 27 servers at peak running. Watch this point (in my case)
4. Files bigger than 4096 MB won't fit, but most rippers can split, so I never cought this issue here.
5. The theoretical limit of FAT32 is 128 GB - 1 MB, FAT32 is really FAT28 (no joke!) (Due to the hardware of this very day, tommorrow?)


Am I right? Is this really a concern of?
Why is Windows X64 far more faster on FAT32? And why is FAT32 'dead'?
Maybe, I don't understand the hype of NTFS. Also NTFS has it's own disadvantages. NTFS is really nice on a U320 harddisk not for IDE or cheap Sata disks. The Raptor will change it all.

I'm serious. My English is far from perfect and I beam up my skills here. If I'm making big faults, please report if you like.

The theoretical limit is much bigger. As you mention, 28 bits are used for cluster addressing. That makes a maximum of 268 million clusters. Clusters can be as big as 32 kB, which gives a maximum of 8 TiB. (The FAT will be 1 GiB on such a beast)

Yes, that's right, my fault. I mean to say : 128 GB is the practical limit

Still Thanx

I don't know if volumes bigger than 128 GB are able to create and such volumesizes are unpractical.

I must say I'm using the Paragon partition Manager defrag methode. It defrags my 25 gig in 34 minutes and it whas a real mess. FAT32 or rather the Microsoft API's lacks file alignment so the partitions are not created by WIndows setup. I created them with PPM.
I'm using 32 Kb units on each volume and it won't freed up so much space I aspected. My average filesize is 3 MB so the wasted space could be maximal 1% of the total space without MFT zone reservation and such things more (these things also freed ip space and far more than 32 KB clusters)

I think you're mistaken. When you already default to the biggest possible clustersize, increasing a FAT volume doesn't add extra overhead/slack. WinXP can't create a FAT partition bigger than 32 GiB, but that limit is artificial.
3th party PM can create FAT32 partitions as big as you want it.

However, there are two ~128 GiB limits which can be confusing you: The LBA28 limit, which limits the number of adressable sectors to 2²⁸, which is 128 GiB. This limit is overcomen by the LBA48 addressing scheme, which is supported by XP SP1 and W2000 SP3. The second limit is the W9x disk tools limit. W9x Scandisk and Defrag cannot handle partitions bigger than ~127 GiB due to a poor memory management.
Of course none of these issues can hurt an XP64 system.

Some other remarks:

This will be the journaling system.


Why *should* it be infected? Infected with what?

Thanx in advance, that's exactly what confused me. You couldn't wrote it better.
But,,

If all files are available, every good backdoor can upload it because there no rightsmanagement.
Indeed I use a thirt party partition manager and a real good one Many corporations here swere by it.

I've a third harddisk with Vista present and this one already resides on the disk so I plug it in yesterday late.
I'm in a doubt which suits better for me. Vista or X64. Both works perfectly

Important note/question: FAT32 is always FAT28 because the last four bits were reserved by Microsoft?? Not a joke I've red ?
Found the answer: http://www.microsoft.com/technet/prodtechn...i.mspx?mfr=true

I'm right, that's the kind of limitation.

Link: http://support.microsoft.com/kb/184006/

Link: http://support.microsoft.com/default.aspx?...kb;en-us;310525

You're also right , but there's still another limit I thought it is the reason of the 128 GB limit of FAT32. No Offense, you looking it just from another side, but LBA28 is also a limitation, but also for NTFS then, so I mean pure FAT32

One question yet:

If FAT32 becomes a FAT28 on Windows 200 it does on XP even so. The fastfat driver have still the same layout on every NT machine including Vista. Why? On every FAT32 volume you can install Windows 2000 so that limit rises everywere when deploying FAT32.

Is this due since Windows 2000 or Windows 2000 specific? If it is a specific matter, than XP could never be installed at FAT32 due to the backward compatibility??!

I'm sorry, I really don't know what you're talking about.
FAT32 has a 32 bits address for each cluster in the FAT. For W2000 the four most significant bits are reserved. Maybe they weren't for W9x, but that doesn't matter, since these bits will only be used on partitions bigger than 8 TiB, which is impossible on W9x, because it can't handle disks bigger than 2 TiB due to its 32 bit sector counter.
In this document, page 16, you can see that those bits are masked away anyway.

I cannot follow your argumentation that due to the use of only 28 bits XP cannot be installed on FAT32.


To get a backdoor it first has to be installed. The first defence firewall/virusscanner/browser safety will catch away most menaces. When the filesystem had to protect against these malware the war had been lost yet, since many (most?) XP systems are run in an administrator account, switching off all security.
Further, the files which are protect by the filesystem are in most cases the OS itself. Why would a trojan upload these? You can copy them from each XP cd. The most interesting files are personal documents, which are available anyway, because the user is logged in.