RAID level features

What is RAID?
1. How to implement RAID
* An external disk array
* In-Connection raid
* Software RAID
2. RAID level
RAID-0 (Strip, strip Reel)
RAID-1 (Mirror, mirrored volume)
RAID-4
RAID-5
RAID-6
RAID-10
RAID-01
RAID-50
JBOD
Summarize

What is RAID?

Raid was originally called "redundant Arrays of inexpensive Disks" "fault-tolerant low-cost disk array" and later changed to "redundant Arrays of independent Disks" "Fault-tolerant Independent disk array", which is used To improve the overall disk system IO capability, redundancy, and free space. A disk array is composed of several disks, grouped into a larger disk group, providing different capabilities depending on the raid level selected.

1. RAID Implementation Method * External disk array

The expansion card provides the adaptability, the expansion card is called the adapter adapter;

* In-Connection raid

Motherboard Integrated RAID Controller

* Software RAID2. RAID level RAID-0 (Strip, stripe volume)

is to distribute the data evenly on the individual disks of the array in the form of strips;

Characteristics:
Number of disks: Minimum of 2;
Advantages: Read, write performance improvement, there is no verification, not too much CPU resources, design, use and configuration is relatively simple;
Disadvantage: No redundancy ability, can not be used for high data security requirements of the environment;
Free space: N (number of hard drives) *min (S1,s2 ... Depending on the space of the smallest hard disk)
Applicable areas: Video generation and editing, graphics editing, and other needs of large transmission bandwidth operation;

RAID-1 (Mirror, mirrored volume)

Redundant copies of data on a virtual disk, on a member disk, in the form of a mirrored redundancy

Characteristics:
Number of disks: Minimum 2, 2n, n greater than or equal to 1
Advantages: Read performance improvement, write performance decreased slightly, with 100% data redundancy, provide the highest data security guarantee theoretically can achieve twice times the reading efficiency design and use is simpler;
Disadvantage: high overhead, space utilization is only 50%, in writing performance is not very important;
Have redundancy capability
Free space: 1*min (S1,s2 ... Determined by the minimum hard disk space)
Applicable field: Finance, finance and other high-availability, high-security data storage environment;

RAID-4

Data are stored in sequence on multiple hard disks, a separate hard disk to do redundant backup, fault tolerance has been raised. However, independent hard disk access pressure is large, there are performance bottlenecks. 3rd disk storage Verification code.

Advantages: Balanced consideration of performance and redundancy.
Cons: Fixed redundant disks become disk array I/O bottlenecks.
Minimum 3 drives required;
Data is stored in 2 hard disk, and then the 3rd hard disk to store the data check code;
The check code is the value of the chunk block in 2 hard disk which is different or calculated by the bitwise;
1 of the hard drive does not affect the file data read and write operations, the data can be recovered, but it is somewhat slow, even if the bad 1 hard disk still continue to work online, called the downgrade mode, the data is not guaranteed, the risk is greater; so to replace the bad hard disk with a new hard drive, tentative business, with 2 free disk You can restore data to a new hard disk by bit-checking, and when all the data is restored to the new hard disk, it will continue to work, but in case there is a risk in the recovery process;
RAID4 also has an inherent disadvantage: the use of a single disk as a store check code, regardless of the front of the disk access data, the calibration disk must be accessed, that is, centralized check code check disk access pressure is too large, it is easy to cause performance bottlenecks; Therefore, early detection of bad disk damage, can be replaced as early as possible 1 new hard drives can be used as spare disks.
XOR operation, store check code:
For example: 1101,0110 bitwise check, CHECKSUM code: 1011

RAID-5

Using the independent Access array method, the calibration information is distributed evenly to each disk of the array;
Compared to the RAID-4 to store the check code on a hard disk, and the RAID-5 is the 3 disk rotation for the storage check code. The order of the left symmetry that the check code holds each disk is first in the first 2 pieces of inventory data, the 3rd inventory check code, and so on, and so on, right symmetry opposite

Characteristics:
Number of disks: Minimum of 3
Advantages: High reading performance, medium write performance, check the distribution of information access to avoid the bottleneck of writing operations;
Disadvantage: Controller design is complex, the process of disk reconstruction is more complicated;
Free space: (N-1) *min (S1,S2, ... The smallest of these spaces)
Fault tolerance: 1 disks
Applicable domain: File server, email server, Web server and other environment, database application;

RAID-6

A verification code is added on the basis of RAID-5, which greatly improves the redundancy performance. / two disk to do check disk, check the code to save 2 times, rarely used /

RAID-10

RAID10 combined with RAID1 and RAID0, mirror first, then stripe

Characteristics:
Number of disks: Minimum 4, 2n, n greater than or equal to 2;
Advantages: Read performance is very high, write performance is good, data security is good, allow simultaneous n disk failure;
Disadvantage: The utilization rate is only 50%, the cost is big;
Free space: N*min (S1,S2, ... where the minimum space)/2;
Fault tolerance: Each group of images can only be broken one piece;
application area: Used in database applications requiring high availability and high security;

RAID-01

First divided into two groups to make RAID-0, and then the composition of the RAID-0 made into RAID-1, do not conform to the common methods, each group has a bad hard disk is likely to be large;

RAID-50

Is the combination of RAID5 and RAID0, first realize RAID5, and then strip, (first do RAID-5 in RAID-0, at least 6 disk, each group allowed bad 1 disk, space utilization flexibility), RAID7 (a company's private technology, is actually a file server)

Characteristics:
Number of disks: minimum of 6;
Advantages: Better reading performance than RAID5, less than the same capacity RAID5 reconstruction time, can allow n disks to fail at the same time;
Disadvantages: Complex design, difficult to achieve; the failure of two disks in the same RAID5 group can cause the entire array to fail;
Applicable domain: Large database server, application server, file server and other applications;

JBOD

Just a Bunch ofdisks
Combine multiple disk space into one large contiguous space for use;
Free space: sum (s1+s2+, ... The sum of disk space)

Summarize

Comparison of common RAID levels: Raid-0,raid-1,raid-5,raid-10,raid-50,jbod
RAID-0 the best performance;
RAID-1 redundancy is highest;
RAID-1 and RAID-10 have the highest overhead in the same usable capacity;

From for notes (Wiz)

RAID level features