Brief Explanation About RAID And Array Controllers
When it comes to a server, calculating the necessity of the storage is a must. Especially when you are considering what kind of roles of your server will be. Every role of your server should be set along with how much space you would reserve. For example, if you willing to create a file server role, you should consider how much user that will save their files into the server, how big the file is, and you should also consider how much space will be created.
Along with that, the risk of losing data is should be considered too. Is this file server should handle these files and make a backup, just in case there's a failure? or, is it should be only use to save files without backing up? Well, the choice is all yours. But before picking which one is good, you should understand the fundamental of server storage redundancy.
Where To Begin With?
To make it coherent for your understanding, let's comprehend that most server nowadays comes with a device called an array controller. Array controller is a device that handles multiple storage (hard drive or SSD) inside your server. It is able and could combine multiple storage into one logical partiton. For example, you have 2 hard drive which has 1 TB volume each. These array controller could make these two hard drive readable as 2 TB hard drive. It is possible because array controller created something called "array".
What is Array?
Array is actually block of storage. Array consist of multiple storage (hard drive or SSD) combined in one partition or more. This block of storage has a technology called RAID.
What in heck is RAID? Is it Related to Raid: Shadow Legend Game?
Nope. RAID is a short term of "Redundant Array Of Inexpensive Disk" or "Redundant Array of Independent Disk". This is the real "magic" behind those combining storages inside an array. RAID is a technology that able to combine the disk for the necessity of data redundancy or performance improvement, or even both. RAID has certain types that could later be useful to fullfil the role of server. It's RAID 0, RAID 1, RAID 10, RAID 5, RAID 6. Below is the explanation in text, but if you willing to watch the explanation with videos, you can head to these video explanation between RAID by Powercert Animated Videos on Youtube.
|RAID illustration By Powercert Animated Video|
RAID 0 is combination of 2 hard drive that is allocating it's space to save information by stripping it into the both hard drive. For example, you have 2 hard drive with 1TB volume each. This RAID 0 will combine those 2 disk, make it as 2 TB volume, and write everything into it. This kind of redundancy is not fault tolerant, because if one of your hard drive fails, all of your data will be lost. That's because one information was spread into those two drive. Once the one drive are needed to be read, it also need another to do so. RAID 0 is never a good-go to use as big data host because it doesn't create a copy. The one and only reason to use RAID 0 is for it's fastest read and write process. RAID 0 are best to use as a front-end webserver host.
RAID 1 use the minimum combination of 2 hard drive that is use as data storage and a mirror. For example, you have 2 hard drive with 1TB volume each. This RAID 0 will write your information and data to a single hard drive, and make a copy to another. In this case, your first 1 TB will be used as write-place and your another 1TB will be used as a copy. This is the most common RAID that are used in small to medium company because it provide the fault tolerance. If one of those 1 TB failure, you just need to unplug the fail ones, and replace it with brand new. The array controller will rebuild the copy and you really good to go. With that 2 hard drive, you can only use 1 hard disk only.
RAID 10 is actually combination between RAID 1 and RAID 0. RAID 10 need minimum 4 hard disk to be able to work. The benefit of this redundancy is, the fast transfer-rate of RAID 0 and the fault tolerance of RAID 1 both combined. For example, you have 4 hdd which has 1 TB volume each. This 4 hard disk will be split into 2TB x 2TB. The first 2 will be combined and used as write-place, and another combined 2 will be used as mirror.
RAID 5 use the minimum combination of 3 hard drive that all of it are use as data storage. RAID 5 use the same technology as RAID 0 by stripping every information and data into each hard drive separately. But the difference is, RAID 5 create a strip called parity. Parity is a strip where all mirror are saved. Parity are created in every hard drive that is allocated inside an array. For example, you have 3 hdd which has 1 TB volume each. All of the information will be write into those 3 hard drive, but in each drive will also created a parity. in this example, parity will cost you 1 TB of space as a backup or mirror. If one of those 1 TB failure, you just need to unplug the fail ones, and replace it with brand new, and the array controller will collecting parity from every disk to the new ones. RAID 5 could only handle 1 disk failure at a time. So if two disk fails at the same time, the data will be lost.
RAID 6 use the minimum combination of 4 hard drive. RAID 6 is almost same as RAID 5, but RAID 6 write double parity strip inside every hard drive assigned. For example, you have 4 hard drive that every one of it is 1TB. Every information will be write into these 4 hard drive, and parity will be stripped twice in each disk for the sake of preventing two fails disk at the same time. RAID 5 and 6 have the same writing time but fault tolerance is much bigger with RAID 6 than 5.