Breaking the 2.1 Gigabyte Barrier
BREAKING THE 2.1 GIGABYTE BARRIER:
USING LARGE-CAPACITY DRIVES WITH DOS,
WINDOWS AND WINDOWS 95
A Quantum White Paper
As hard disk drives for desktop PCs move to multi-gigabyte (GB) capacities, they have outpaced the ability of the
most commonly used PC operating systems to manage them as a single disk partition. In order to take full advantage of
todayís higher storage capacities, users of Windows® 95, Windows 3.x or DOS-based PCs must
partition hard drives that exceed 2.1GB into multiple logical drives of lesser capacity.
Quantum's newest desktop hard drives push past the 2.1GB barrier
Large databases, more sophisticated application software, multimedia files and other graphics-intensive data: they are
continuing to drive the need for larger capacity/lower-cost-per-megabyte hard disk drives. The File Allocation Table
(FAT) portion of many PC operating systems has a built-in 2.1GB addressing limitation. Several of Quantum's newest
drives for desktop PC systems eclipse this capacity point, creating a problem if the entire drive is to be used as a single
partition. These new drives include:
- Quantum Bigfoot™ 2.5-2550MB
- Quantum Fireball 2.1 TM-2160MB
- Quantum Fireball 3.2 TM- 3240 MB
Through partitioning, the drives listed above and future high-capacity drives can circumvent the FAT file systemís
Limits in the file allocation system
All data stored on hard disk drives relies on some type of file allocation or addressing system. It is, however, important
to understand why the 2.1GB barrier exists and what choices PC users have.
In the world of desktop PCs the most common file system is the FAT system used by all Windows and DOS-based
machines. However, the current FAT file system has an inherent addressing limitation that prevents it from supporting
drive partitions larger than 2.1GB. This limitation lies within the FAT file system as used by Windows and DOS.
Advanced file systems, such as those used by Windows NT, OS/2 and the Macintosh operating system, do not impose
Under DOS 6.x, Windows 3.x and Windows 95, disks with capacities of more than 2.1GB must be partitioned into smaller
logical units in order to be fully accessible by Windows. This is not necessarily a problem since partitioning can
actually enhance the performance of drives relying on the FAT file system. In fact, partitions have historically been
used to divide drives into multiple, logical sections ó often to support multiple operating systems on the same drive.
Architectural overview and historical limitations
Four basic systems components are involved with accessing hard disk drives:
- Disk drive - physical storage device
- BIOS (basic input/output system) - software for providing access to the storage device
- File system - software that manages file storage and retrieval
- Operating system - software that provides overall management of applications and resources
Disk capacity addressing limitations, regardless of the physical capacity of the disk drive itself, are nothing new for
DOS-based PCs. Prior to 1989, for example, there was a 32 megabyte partition limitation. Partition size on disks drives
associated with DOS machines was limited to 32 megabytes due to the FAT file systemís fixed 512 byte cluster sizes.
This was solved by a change to the operating system to support the use of multiple partitions on a single drive. Also,
beginning in 1985, the 16-bit FAT file system, which could support larger partitions through the use of increased cluster
sizes, was implemented. (see section below, File Systems).
The current limitation, the 2.1 GB partition barrier, is once again a FAT file system issue. The FAT file system is the most widely employed and most venerable of the PC file systems, having been originally conceived in 1976, long before the advent of high- capacity disk drives.
Essentially, file systems are responsible for how operating systems organize data on a disk or other storage medium.
From their inception, DOS- and Windows-based PC's have, for the most part, used the FAT file system. Windows 95
uses a FAT variant called VFAT, which supports longer file names.
The OS/2 operating system typically uses another file system altogether, called the High Performance File System
(HPFS). Windows NT-based workstations and servers employ the NT File System (NTFS). Both these 32-bit operating
systems also support the 16-bit FAT system. However, DOS, Windows and Windows 95 cannot recognize files created
under HPFS or NTFS.
Why does the FAT file system's 2.1GB barrier exist?
File systems organize disk space in allocation units called clusters. In the FAT (and VFAT) system, the number of clusters can be a maximum of 64K (65,536 bytes). FAT file system clusters vary in size according to the size of the partition. As the partition gets larger, the clusters get larger: up to a maximum 32K (32,768 bytes). Multiplied together, 64K times 32K equals a maximum logical drive size of 2GB (2,147,483,648 bytes).
Partition size Cluster size Maximum number of clusters
|Partition Size||Cluster||Maximun number of cluster|
|Fat Limitation: 32 K x 64 K = 2Gbytes|
What this means for DOS, Windows and Windows 95 users is that their PCís operating system currently cannot
address a disk logical partition larger than 2.1 GB. Larger drives require multiple partitions to provide the use of full
storage capacity. Once again, Windows NT and OS/2 users are not affected if they are not using the FAT file system.
|Operating System||Supported File Systems|
|Windows 95||VFAT (FAT variant)|
|Windows NT||NTFS, HPFS, FAT,VFAT|
|File System||Partition Size Limit|
|FAT||2GB (2,147,483,648 bytes)|
|HPFS||2 Terabyte (2,199,023,255,552 bytes)|
|NTFS||16 Petabyte (18,446,744,073,709,551,616 bytes)|
Solution: disk partitioning
Hard drives larger than 2.1 GB can be partitioned into logical partitions that are each less than 2.1GB. Several software
utilities are available for the task, including the FDISK utility, which is included with the DOS and Windows 95
operating systems. Optimal partitioning depends entirely on user needs, and there are various performance and
efficiency issues to consider. Typical partitions are:
Theoretically, a desktop computer system can support up to 24 partitions per file system: one for each of the letters of
the alphabet, minus A and B which are used to identify floppy drives. Each partition represents a logical drive to the
operating system. A system may have multiple hard drives, but a single partition must be contained on a single physical
device ó it cannot be spread over several drives.
It is important to bear in mind that re-partitioning a drive can permanently destroy all previously stored data on that
drive. Also, for simplicity, it is best to keep the number of partitions low.
Partitioning for performance
Disk drives can be partitioned to increase system performance under the FAT file system. A 2.5GB drive, for example,
may be partitioned as a relatively small segment of 400 megabytes, and used as the primary or C drive. This can be used
for operating system files, including the file swap area, as well as commonly used applications that appear as icons on
The rest of the disk can then make up a secondary partition of 2.1GB for user files, less frequently used applications,
and games (including CD-ROM games that can require hundreds of megabytes of disk space for virtual memory).
A relatively small primary partition best serves the administrative functions of the operating system as it engages in
various background tasks that swap data between system RAM and the hard drive. Reducing the size of the partition
naturally reduces the amount of physical seeking necessary for these tasks because all the files are stored in the same
Also, under the FAT file system, a relatively small partition translates into a relatively small cluster size.
Alternatives and future solutions
The 2.1 GB capacity addressing limitation is entirely a FAT file system issue. While the industry awaits a solution from file system software developers such as Microsoft, users can manage higher capacity drives and tune performance through partitioning. Windows NT or OS/2 are also viable solutions because these operating systems support file systems with size limitations above 2.1GB.
Storage Capacity Notation
There are two methods commonly used to reference the size of memory and storage devices. One is a base-ten method, and the other is a base-two or binary method. While file systems generally report storage capacity in binary terms, it is customary for disk drives to be described under the base-ten system, as is the case in this paper.
|Base-ten (disk drives)||Binary (file system)|
|1 MB = 1,000,000 bytes||1 MB = 1,048,576 bytes|
|1 GB = 1,000,000,000 bytes||1 GB = 1,073,741,824 bytes|
|2 GB = 2,000,000,000 bytes||2 GB = 2,147,483,648 bytes|