Buying Hard Drives

The hard drive is one of the most popular upgrades for the PC. Like RAM, you can never have enough hard drive storage space. It always seem to run out before you realize it. Maybe it's all that porn and bootleg anime you have been downloading. So, should you just run out and buy the hard drive with the largest capacity that you can afford? Or is there more to it than just storage space?

This guide is designed to give you the low-down on hard drives and what makes one better than the other. I will go through the parameters and metrics experts use to decide on the best hard drive for the job. After reading this guide, you will be better armed to select the best hard drive for your needs. Let's begin!

Performance

Unlike processors, hard drive performance isn't quite as simple to determine. Certainly you cannot count on simple benchmarks that use some kind of unified scoring system. Although the general tips on choosing the right processor benchmarks still apply here, we'll have to be a bit more liberal when it comes to hard drives. But before we go into the topic of testing the hard drive's performance, let's go through the performance metrics of a hard drive.

Read Speed – This is a measure of how fast a hard drive can deliver requested data (from either buffer memory or hard drive platter) to the hard drive controller on the motherboard. Obviously, the higher the read speed, the better. Do note that because this read speed is influenced by the hard drive's SDRAM buffer and controller, it will not give you an accurate reading of the drive's physical performance. However, it is a good indicator of the drive's real world read performance.

Write Speed – This is the reverse of read speed. Instead of measuring the speed of data transfer from the hard drive to the hard drive controller, it is the measure of how fast a hard drive can complete writes from the hard drive controller. Again, its result is influenced by the hard drive buffer and buffer controller. So, it is a good indicator of a hard drive's real world write performance.

Internal Read Speed – To test this, the buffer is disabled and data is read directly from the hard drive platters. This allows you to isolate the drive's physical read performance. The internal read speed is often recorded as a graph as the read speed on the outer tracks are much higher than the read speed on the inner tracks. The graph will often show a step-ladder reduction in read speed as the platters are read from outside in. This is because of the Multiple Zone Recording (MZR) technique used these days.

Although benchmarks will usually list both maximum and minimum internal read speeds, manufacturers often quote the hard drive's maximum internal read speed (for obvious reasons!) which is recorded at the outermost track. This is also known as the maximum sustained read speed because it indicates the maximum read speed the hard drive can sustain after its buffer memory has been exhausted. It is a good indicator of the drive's read performance for files that are much larger than the buffer size.

Internal Write Speed – This is the reverse of internal read speed. It measures the time it takes for the hard drive to directly write data to the platters, without the involvement of the buffer. Although there is always a maximum and minimum internal write speed, usually only the maximum internal write speed is quoted, often as maximum sustained write speed. It is a good indicator of the drive's write performance for files that are much larger than the buffer size.

Burst Speed – This is a measure of how fast data can be transferred from the buffer to the hard drive controller. In other words, it is an indication of the speed of the interface as well as the buffer and its controller. The higher the burst speed, the better the drive will perform.

Track-to-Track Seek Time – This is the measure of how long it takes to move the read/write heads from one track to an adjacent track. This is the fastest possible seek speed that the read/write heads are capable of. The track-to-track seek speed is given in milliseconds and naturally, the lower the seek speed, the better.

Random Seek Time – This is the average measure of how long it takes the read/write head to seek out random pieces of data . This metric is especially important in server environments where data are often accessed randomly. Measured in milliseconds, lower random seek speeds are always better.

Rotational Latency – This is the average measure of how long it takes for the platter to rotate a full revolution. This is entirely determined by the platter rotation speed, which is also known as the spindle speed. The higher the spindle speed, the lower the rotational latency. Measured in milliseconds, a lower latency is always preferable.

Access Time – This denotes the average time it takes for a hard drive to receive a command, move the heads to the appropriate position and start the read/write process. As such, it is a combination of the various metrics above. This can be roughly summarized using the simplified formula of Access Time = Command Overhead + Seek Time + Latency. Again, the lower the access time, the better the hard drive's performance.