If you've been working with the computer long enough, you're bound to have the fun of pulling out a disc or even a 3.5-inch floppy disk during a sweep. These long-obsolete storage media, square or round, must still hold an interesting record of "early human taming data". Yes, the need for data to be backed up in a timely manner is also an unbreakable topic. The American online community Reddit even launched an initiative to recommend March 31, the day before April Fool's Day, as "World Backup Day" to remind everyone that only fools forget to back up data.
Living in the information age, we are actually dealing with massive amounts of data every day, at least the operators are texting me every day to remind me of the remaining traffic of the package. Maybe you will say that data is not in the computer and mobile phone? But with just a cup of milk tea that you accidentally poured, the important documents you usually don't care about will be quickly destroyed. And the best way for us to escape this anxiety is not to ignore it, but to back up the data in time.
Simple data backup, we usually use it. As small as a USB stick, you can double the safety factor of your home electronic photo album. But as we need to store more and more files, or save and read important files at any time in our work, I believe many people have chosen cloud disks, NAS, and even small servers. Of course, there is also a DAS (direct-attached storage) that is more suitable for direct connection for small and medium-sized enterprises. But in fact, in the end, the storage backup technology they use is much the same, at least the word you should not be unfamiliar with - RAID.
In 1988, Professor D. A. Patterson of the University of California first proposed the concept of RAID, and at first, RAID (Redundant Array of Inexpensive Disks) was officially called "cheap redundant disk array". Such a strange name was born, mainly because at that time, large-capacity disks were very expensive. Therefore, how to use a combination of multiple disks with small capacity and easy cost control to obtain the same capacity, performance and reliability as large-capacity hard disks has become a very critical issue.
As soon as the new idea of RAID was introduced, it was absorbed by the industry, and a number of new storage technologies with high performance and high reliability were quickly developed, and they have been widely used today. Nowadays, the meaning of "I" in RAID is no longer "Inexpensive cheap", but has become "Independent independent"; RAID Tiantuan has also changed from the original D. The "five small strengths" of RAID 1 to RAID 5 defined in Professor Patterson's paper extend RAID 0, RAID 6, RAID 7, and even RAID 10/01, RAID 50, RAID 53, and RAID 100. So what do these RAIDs, which sound like they are constantly "upgrading", have to do with each other? How exactly do different RAIDs help us?
First, let's clarify the concept. RAID is actually a collective term for a class of "multi-disk management technology", which we generally refer to as "disk array". Disk arrays utilize physical storage space partitions to back up redundant information for our data. Once a disk or access path fails, redundant information can quickly help users recover data saved in failed storage. This is the most common use of RAID - daily backup and recovery of data. Different RAID levels are more different from each other in different ways than in the difference between advantages and disadvantages. Usually the seven levels of RAID 0 to RAID 6 are defaulted to standard RAID levels because they contain the most basic and commonly used RAID configurations.
RAID 0 and RAID 1
First of all, there is a famous bastard in the RAID family - RAID 0, and even RAID 0 is not a real RAID method. RAID 0 is the simplest to implement, because it actually combines different hard disks into a "big hard disk", such as a 2TB hard disk and a 4TB hard disk with RAID 0. The answer is that it will become a 6TB "big hard disk", which has increased in capacity but does not have any backup function itself. Suitable for applications where hard disks with the same performance level are mixed and do not require data security. As for why RAID 0 can be successfully mixed into it, it is also thanks to the fact that it can increase the read and write speed of the hard disk array, and theoretically can achieve the effect of direct × N. Therefore, RAID 0 has also become a very common way of RAID.
RAID 1 is also very simple, the data in RAID 1 is in the way of "1 + 1", the two hard disks are backed up at this time, one has a problem, the other can be top at any time, at this time we only need to replace the fault disk. The advantage of RAID 1 is that the way to group arrays is simple, the disadvantage is that it is not capacity friendly. Because two 1TB hard drives make up RAID 1, the actual free space is only 1TB. It seems very unscientific, but in fact, RAID 1 is indeed a very safe backup method, suitable for servers, databases and other applications with high security requirements. However, compared with RAID 0, the disadvantage of RAID 1 is that it does not improve the read and write performance of the disk array.
RAID 2
RAID 2 is a bit more advanced than the two brothers. It uses "Heming Code" to achieve data verification redundancy. The advantage of Haiming code is that it has the ability to correct errors, and the complexity is not high. The disadvantage is that this error correction method itself is not very clever, because the data is stored bitwise in RAID 2, 4 bits of data require 4 data hard disks, and 16 data files need 16 data hard disks. Obviously, this combination is very inflexible, so we rarely see RAID 2 in everyday use as a form of disk array.
RAID 3 and RAID 5
RAID 3 and RAID 5 are obviously more flexible than RAID 2's "one is one, two is two" disk array approach, because they only need at least 3 hard drives to achieve whatever kind of data. Among them, RAID 3 has a special hard disk as a check disk (RAID 2 also uses the check disk) to store a separate parity information. When one of the hard disks has a problem, all the data in the failed disk can be recovered by using another normally working data disk and the calculation and reconstruction of the verification disk. Therefore, in theory, RAID 3 is very suitable for home application scenarios that are more sensitive to the price of storage media, and can be used to store some video materials, streaming media materials, etc. Daily, and video and other visual creative workers with huge amounts of materials are recommended to directly choose RAID 3 as a backup of daily data.
The difference between RAID 5 and RAID 3 is that RAID 5 is an array method that does not have a dedicated check hard disk. The check data for a RAID 5 array is randomly distributed across all disks. Raid 5 advantages and disadvantages are prominent, the advantage is that it has very good flexibility and scalability - RAID 5 can increase the number of array hard disks at any time, and even RAID 5 can also simulate RAID 3 mode of operation, and in theory RAID 5 can achieve (N-1) read speed increase, so it is very suitable for storage arrays used as high bitrate streaming; the disadvantage is that RAID 5 is a compromise between RAID 0 and RAID 1. The utilization of the hard disk is comparable to THAT 3, which is N-1.
RAID 4
Raid 4 principle is the same as RAID 3, but the verification method is slightly different. But it is the difference between big and small and small that makes RAID 4 more restrictive than the first two applications. RAID 4 takes the form of a statistic block using the same XOR check. Even validation can be achieved at the time of reading, that is, it does not affect read performance. However, the write can only be carried out on a single disk, and the check data must be written at the same time, so the write speed of the check disk can easily become the performance bottleneck of the entire array (this is also a problem with RAID 3, but the combination of RAID 3 is more flexible, so the application scenarios are more extensive). Raid 4 is rarely seen in mainstream storage products.
RAID 6
Seeing the small partner here may have a big doubt, the RAID level mentioned earlier, it seems that it can only protect the failure of a single hard disk. Because once two data disks fail at the same time, the entire array will still crash. Then it's RAID 6's turn to play. Raid 6 introduced the concept of double check for the first time, that is, when two data disks fail at the same time due to distortion in the array, the entire array can still ensure data security. At the same time, RAID 6 is also a "hardened level" scheme based on the RAID 5 idea, which can be seen as a level extension of RAID 5. Of course, the price is that RAID 6 requires at least 4 hard disk combinations (N-2 constitutes a backup), which is often suitable for applications that require a very high level of data security.
RAID 10
In addition to the above RAID "seven brothers", there is a very common RAID method - RAID 10, which some people also call RAID 01. Because RAID 10 is actually a combination of RAID 0 and RAID 1. First build a RAID 1 backup of the hard disk in pairs, and then build an extension through RAID 0. Hence the use of efficiency and safety aspects. RAID 10 combines the advantages and disadvantages of RAID 0 and RAID 1, and the hard disk usage rate is only half of the security at the same time, but because the combination method is relatively simple, it is suitable for applications where security and efficiency need to be compromised, such as the financial industry, film creation and other fields. Similarly, the large digital RAIDs mentioned earlier, such as RAID 50, RAID 53, and RAID 100, are also an organic combination of the above seven common RAID methods. But in addition to RAID 10, several other RAID grades are more expensive to implement and are therefore less common.
The following is also a summary of a brainless selection method for everyone:
The pursuit of large capacity, regardless of data security = RAID 0;
Do not pursue large capacity, consider data security = RAID 1;
That is, large capacity, but also data security, large file storage is the mainstay = RAID 3;
That is, large capacity, but also data security, small file storage is the mainstay = RAID 5;
Higher data security requirements = RAID 6.
Knowing the above RAID levels, you will understand why the common NAS is divided into two-, four- and six-plate. Dual-bay bits are only suitable for bulk and RAID 0, RAID 1, and quad-bay bits are in addition to RAID 0 and RAID 1, adding support for RAID 3 and RAID 5, scalability and RAID flexibility, compared to dual-bay bits to improve a level. As for the six-bay position, that is obviously designed for a more secure RAID 6 and larger capacity.
As we all know, the high-performance and high-reliability Western Digital WD Red Series has always been an excellent choice for NAS users. We have also evaluated the WD Red SN700 NVMe SSD as a high-performance SSD, but for users with a large number of I/O-intensive workloads, the large-capacity, high-performance, high-security WD Red SN700 is undoubtedly an excellent choice for building NAS caches.
WD Red SN700 NVMe SSD
See this article for details:
Recently, western digital launched the WD Red Pro 20TB HDD for single-disk mass storage needs, with the help of nine ePMR (Energy Assisted Vertical Magnetic Recording) platters, crammed into an exaggerated 20TB capacity in a 3.5-inch volume. And Western Digital's WD Red Pro 20TB uses Western Digital's OptiNAND technology to deliver higher performance, reliability, and capacity through iNAND UFS embedded flash drives (EFDs). In terms of parameters, the Western Digital WD Red Pro 20TB adopts SATA interface, comes standard with 512MB cache, has a speed of 7200RPM, and has excellent read performance as fast as 268 MB/s. For small partners who pursue the ultimate single-disk capacity, WD Red Pro 20TB is obviously a good choice.
WD Red Pro 20TB HDD