The development of the hard disk access time can not keep up, which is why it has become the bottleneck with the other computer components such as CPU, RAM or graphics card. To achieve high performance, a hard disk must therefore as much as possible always read or write large amounts of data in successive frames.
From the definition of it is clear that two half large plates running at the same speed provide the same amount of data with twice the number of IOPs. This is one reason why server disks are typically not as large as desktop plates as shown by Computer Repair Shop Sydney.
This is achieved, inter alia, that as many operations performed in RAM and the positioning of the data is adapted to the access pattern on the plate. This is mainly a large in-memory cache of the computer that is provided by all modern operating systems. In addition, the hard drive electronics has a cache (as 2012 for plates 1-2 TB usually 32 or 64 MiB), which serves to decouple the interface transfer rate of steady transfer rate of read-write head.
In addition to the use of a cache, there are other software strategies to improve performance. They are particularly effective in multitasking systems, where the hard disk system with several or many read and write requests is confronted simultaneously. It is then most efficient to bring these requirements into a meaningful new order. Control is provided by a disk scheduler.
The simplest principle is pursuing the same strategy as an elevator control. The tracks are first approached in one direction and processed according to the requirements, for example, monotonically increasing track numbers. Only when all these are processed, the movement is reversed and then working toward monotonically decreasing track numbers, etc.
Until about 1990 disks had usually so little cache (0.5 to a maximum of 8 KiByte), (then 8.5 KiByte or 13 KiByte) so much that they could not cache the entire track. Therefore, the data access had to be slowed down or optimized by interleaving. This was not necessary in drives with high quality SCSI or ESDI controller or with the then-emerging IDE drives involved in Computer Repair Shop Sydney.
The SSDs used since circa 2008 (Solid State Drive) have no moving parts and therefore have inherently on much shorter access times. For writes, however, to be expected with longer periods than the reading. Since 2011 there are combined drives – transparent to the computer – implemented as a part of the capacity SSD, which serves as the conventional buffer plate.
No Comments