The--raid of storage reliability Technology

When a cloud project is delivered, it is inevitable to consider what raid the storage disk will take. For example: Our project engineer may recommend that you connect a clone virtual machine system disk Group RAID 10, complete replication virtual machine data disk

Using RAID5 or RAID6 and so on, what exactly is RAID? Why should we use RAID? How do we choose the right RAID format?

1) RAID Concept

RAID is an abbreviated redundant array of independent disk, redundant array of independent disks. It is a data protection technology that combines multiple pieces of separate hard disks (physical hard disks) in different ways.

Form a hard disk group (logical hard disk) that provides higher storage performance and data redundancy than a single hard drive.

2) Advantages of RAID

• Improved transfer rate

RAID significantly increases the data throughput of the storage system by simultaneously storing and reading data on multiple disks. In RAID, you can have many disk drives transfer data simultaneously, and these disk drives are logically a disk

Drive, so using RAID can reach a single disk drive several times, dozens of times times even up to a hundredfold rate.

• Provides fault-tolerant functionality

Ordinary disk drives fail to provide fault tolerance, RAID provides fault tolerance, and raid fault tolerance is built on the hardware fault tolerance of each disk drive, so it provides greater security. In many raid modes, there are

More complete mutual verification/recovery measures, even direct mutual mirror backup, which significantly improve the fault tolerance of the raid system, improve the stability of the system redundancy.

Second, RAID classification

At present the industry recognized standard is RAID0~RAID5. In addition, there are other 6,7,10 and so on.

This article gives a brief description of several commonly used RAID levels:

· RAID0

It divides the data into a certain size and writes it sequentially to the disk in the array. As shown in the following:

　　

Theoretically, a RAID0 system consisting of n disks, which reads and writes performance will be n times the read performance of a single disk, and the storage efficiency of disk space is maximum (100%). Due to the influence of various factors such as bus bandwidth,

The actual lift rate will be lower than the theoretical value. However, the performance of a large amount of data parallel transmission and serial transmission is necessarily greatly improved. RAID0 has one obvious drawback: no data redundancy protection, once the data is corrupted,

will not be recoverable. RAID0 requires at least 2 hard drives to be implemented.

· RAID1

Also known as mirroring (mirror), it writes the data exactly to the working disk and to the mirrored disk, respectively.

　　

The RAID1 system has a disk space utilization of 50%, which has an impact on data write time, but does not have any impact when reading. RAID1 provides very good data protection, and once the working disk fails, the system automatically

The mirrored disk reads data without affecting the user's work. and RAID1 support "hot replacement", that is, the failure of the power of the case to replace the disk, the replacement as long as the data from the mirror disk recovery. RAID1 requires at least 2 hard drives to be implemented. ---even blocks.

· RAID5

A standalone disk structure with distributed parity. Data verification information is evenly dispersed across the array of disks, the array of disks have both data and data validation information, data blocks and corresponding checksum information stored on different disks.

　　

The check bit, or P-bit, is different or calculated by the data of the same band. When a data disk is damaged, the RAID5 system can reconstruct the corrupted data based on other data blocks in the same area and corresponding checksum information. RAID5 requires at least 3 hard drives to be implemented.

· RAID6

A disk structure with two parity checks. Based on RAID5, RAID6 further strengthens the data protection, which is actually an extended RAID5 level.

The data redundancy performance of RAID6 is quite good. However, due to the addition of a check, the efficiency of writing is worse than RAID5. Moreover, the design of the control system is more complicated, and the second block also reduces the effective storage space.

RAID6 requires at least 4 hard drives to be implemented.

· RAID10

is a combination of RAID0 and RAID1. The structure of the RAID10 is very simple, first creating 2 independent RAID1, and then the two separate RAID0 to form a RAID0, when the data is written to this logical raid, the data is written in an orderly two RAID1.

　　

The RAID10 performs an array with RAID0 to RAID1 as a data protection array. With the same fault tolerance as RAID1, the overhead for fault-tolerant processing is essentially the same as for a single mirroring operation, with high I/O bandwidth due to the use of RAID0 as the execution level.

RAID10 requires at least 4 hard drives to be implemented.

Usage Scenarios

· RAID0 Usage Scenarios

RAID0 does not provide fault tolerance, but it has a high read performance. Therefore, RAID0 applications are used in situations where the read performance requirements are high, but the data stored is non-critical.

· RAID1 Usage Scenarios

RAID1 provides very good data protection, and once the working disk fails, the system automatically reads the data from the mirrored disk and supports "hot swap", so RAID1 is used for applications where data protection is of paramount importance.

· Usage scenarios for RAID5 and RAID6

RAID5 is a storage solution that combines storage performance, data security, and storage costs. RAID5 can provide data security for the system, but the level of protection is lower than RAID1 and disk space utilization is higher than RAID1.

RAID6 two independent parity systems using different algorithms, the reliability of the data is higher than the RAID5, even if the two disks fail at the same time will not affect the use of data.

Therefore, for the data to save the required level of requirements is not very high, we only need to use the regular RAID5. For now, for those organizations with high levels of data security, such as data centers, information centres,

It is still necessary to use RAID6 to protect data.

· RAID10 Usage Scenarios

RAID10 is ideal for applications such as database storage servers that require high performance, high fault tolerance, but little capacity requirements.

In the actual project, we need to consider the security requirements of the data storage, price requirements and other factors to choose a reasonable raid format.

The--raid of storage reliability Technology