About a Case for Redundant Arrays of Inexpensive Disks(RAID)

A Case for Redundant Arrays of Inexpensive Disks(RAID)

This paper describes a technique that has been widely used in storage to get most out of currently existing inexpensive, or independent disks, that is RAID, short for Redundant Arrays of Inexpensive Disks. The main knowledge I have get when finish reading this paper is some other things about storage via searching for materials about RAID. Now I will conclude all of what I have learned.

To manipulate data on disks, the operation may behave as follows:

 

     1. The user double click the document file in order to read it.

     2. The operation system catches this double-click event, and send a open(read) command to the file.

     3. The file system searches the file through its directory tree, and find the place where the file really residents, then send a open command to the software representation of the file.

     4. The kernel now let the disk driver(it may work as a module) to handle this command, the software disk driver continue send open command to the specific real disk to get the file data that stored on hard disk.

     5. The disk controller(hard drive controller) gets the signal transferred via bus, then send some control signals that indicates to read data to the hard disk.

 

The disk controller is the component which can enable the CPU to communicate with a hard disk, floppy disk or other kind of disk drive. The main technology of the disk controller to interact with hard disk drive are ATA, SCSI, IDE and etc. Note that the disk driver in operating system is different from hard disk driver. The former is a module in OS, the latter is disk storage.

And disk controller is different from what is referred as disk array controller. A disk controller receives the signals on bus, and transfers them to the disk. While a disk array controller, which would been discussed later in more details, is a device which manages the physical disk drives and presents them to the computer as logical units. It can provides disk cache when doing RAID, and it is sometimes referred to as RAID controller.

The disk array controller performs as an interface between hard disk and the processing module, so the end that communicates with disk may be implemented using ATA, SCSI, SATA or etc, but the end that communicates with processing module can be either the protocol above or even network, since the processing module may be another node in a network other than CPU(via bus).

Given these background, we can better understand the RAID.

RAID, according to authors opinion, is a way that offers a cost effective option to meet the challenge of exponential growth in the processor and memory speeds. By presenting several disks as an array, yet logically regard them as a single disk, the throughput of the disk can be larger than original, and the fault tolerance also improved.

Moreover, the RAID techniques nowadays can be divided into software RAID and hardware RAID. Software RAID is that configuring several disk as a logical volume to offer the function to operating system. The hardware RAID is that, using RAID card to organize the disks together.

I believe the main idea of RAID is simple, the details of different RAID levels is also very familiar for us. The author discusses them carefully using most space of the paper, and the description is very clear and easy to read. The details of RAID level can also be easily found on Internet, so I will not describe them again here.

Note that the disk array is manipulated by disk array controller, so if you want to do software RAID, you may implement a controller to do the right thing you expected.

The RAID most used are 0, 1, 5, 6.

 

more info:http://zh.wikipedia.org/wiki/RAID

转载于:https://www.cnblogs.com/kalzium/p/3466024.html