Partitioning

Before you can use any hard drive, it must be set up so that your operating system can recognize it and use it. In Windows and most other operating systems, the first step in this process is called “partitioning” (OK, I lied – the first step is when you physically install the hard drive in the computer, but that’s another issue).

Before going any farther, there are some terms that you need to understand:

Sector: the smallest space on a disk. It’s almost always 512 bytes in size (a byte, for those who are new to this, is eight 1s and/or 0s).

Track: tracks are the way that sectors are physically organized. If you think of them as concentric circles that get progressively larger as you move away from the center of the disk platter, you won’t be too far off. They hold a varying number of sectors, depending on how far from the center of the platter they are.

In hard drives manufactured in the past 10 or 15 years (maybe longer), writing the tracks and sectors is done at the factory and the end-user cannot change or rewrite them without special equipment and software. The track and sector information comprise two-thirds of the physical arrangement of the data on the disk platter. The last part is which side of which platter the data is stored on. So you can immediately find any piece of data if you know the side, track and sector where the data is stored. Unfortunately, operating systems don’t use the physical location to find data. Instead, they use a logical location.

Partition: the largest logical division of a hard disk. You must create at least one partition on a hard disk before an operating system can use it.

Cluster or “File Allocation Unit”: the smallest logical organization unit of the disk. Clusters are composed of sectors and will vary in size, depending on the file system used. They cannot be smaller than 512 bytes (the size of a sector) and will seldom be larger than 64 kilobytes.

In the Windows environment, there are two types of partitions: Primary and Extended. A primary partition is bootable and can contain only one logical drive, while an extended partition is not bootable, but can contain several logical drives. All Microsoft operating systems can recognize up to four partitions on a single hard disk, but a disk cannot contain more than one extended partition.

Before partitioning, you need to decide which file system you are going to use. There are size limitations for all of them.

FAT16: this is the original DOS and Win95 file system. Each cluster is identified by a 16-bit address (hence the “16”). All Microsoft operating systems can recognize and use this file system, but you are very limited in how much space you can allot to a FAT16 partition. It can be no larger than 2GB if you are using a Win9x OS (Windows 95/98/ME) and no larger than 4GB if you are using a WinNT OS (WinNT/2000/XP). The other problem with FAT16 partitions is that they waste a tremendous amount of hard drive space. At 2GB, the clusters of a FAT16 partition are 64 kilobytes. Since a cluster can only hold a single file or piece of a file, this means that for file pieces smaller than 64K, the remaining space is wasted. The space occupied by this article as a plain text file is about 6 kilobytes. On a FAT16 partition, I would have to use a 64 kilobyte cluster to store it, which would mean 58 kilobytes of unused space in that cluster.

FAT32: this is the file system used by Windows 98/ME. Each cluster is identified by a 32-bit address and can be used on partitions larger than 2GB. Windows 2000, Windows XP and some of the very late versions of Windows 95 can recognize and use it; Windows NT and DOS cannot. This file system does not waste as much space as FAT16 if your partition is smaller than 32GB. After 32GB, it’s as inefficient as FAT16, if not more so, and cannot be used on partitions larger than 137GB.

NTFS (NT File System): this is the file system used by all of the Windows NT operating systems. Each cluster is identified by a 64-bit address, so the theoretical maximum size of any partition is several million terabytes (a terabyte is a thousand gigabytes). In practical terms, Microsoft doesn’t recommend partitions larger than about a terabyte. Considering that there are no commercially available hard drives that large (yet), this isn’t an issue. Each cluster is no larger than 4KB, so it uses disk space more efficiently. The downside of NTFS is that the Windows 9x and DOS operating systems cannot read it. In today’s environment where most personal computers are probably running some version of Windows XP, this is not much of a downside, but it is a concern for those who have multiple operating systems on their computer(s).

There are a few more limitations that you need to be aware of before you head off to create partitions on your new disk. The first is that if you’re using a standard IDE drive controller (IDE is actually a very old standard and current drives are ATA, but it’s a distinction without a difference for most people), you cannot have more than two drives on a single controller (most motherboards have two controllers – use one for hard drives and the other for optical drives). If your motherboard supports Serial ATA (SATA), you use one controller per drive and there are usually two SATA controllers on the motherboard itself. Anything beyond those limits will require you to install an expansion card before you can install additional drives.

The second limitation is that the built-in partitioning tools that come with your Microsoft OS _WILL NOT_ allow you to dynamically change any partitions. If there is data on the drive that you want to partition, it will be destroyed in the process. So make sure that you have a full, working backup of all the data on the drive you want to change. There are third-party utilities, such as Partition Magic from Symantec, that allow you to dynamically resize partitions without data loss, but you’ll have to purchase them separately. But a good backup is a good thing to have regardless.