I keep hearing from people in tech communities that keeping things in RAID is somehow better than just slapping it on a hard drive backup, but… why? What even is RAID storage and why is it better?
I spent last night looking up RAID servers on Amazon to get me an answer but I left with more questions: Do all the hard drives in RAID need to be the same size? My entire family switched their laptops due to the work from home orders, so I have a bunch of unused SSDs and HDDs lying around the house, can I just throw those in a RAID server and have it work?
Can I install an operating system on a RAID server? Would it run faster/slower? If I have a MacBook and can install Windows on my RAID server, would it work a bit like bootcamp? What happens when I’m not connected to it?
I HAVE SO MANY QUESTIONS!
EDIT: I think I get it now: It’s a system that’s automatically backing itself up to… itself essentially. You do this by creating redundancies within the RAID system (multiple SSDs). If that’s not it, please tell me.
So final question: can I use a RAID server NOT in RAID? As in can I put multiple hard drives in a server, then plug the server in to my Mac and have it appear as several individual drives? Almost like a HDD mounting system on steroids.
In: Technology
RAID stands for “Redundant Array of Independent Drives” (or sometimes “Inexpensive” instead).
The fact is parts in computers do fail. Hard drives so much more so because they’re often spinning and subject to incredibly tight tolerances so mechanical failure is absolutely a thing. RAID is designed to address this issue by using many disks to cover for each other at the cost that you lose storage capacity. If you need a system that can survive a disk failure RAID is essential. The alternative is to reinstall all the software and restore from backups which can take hours or days during which the computer is effectively dead. With RAID it can soldier on in spite of a dead hard drive.
Various schemes exist, with the most common being:
* RAID-1 – two or more hard drives are kept exactly identical
* RAID-5 – made of many hard drives, one hard drive’s worth of capacity in a group is lost, but any one hard drive can die and the system keeps going. Uses math to calculate missing data from remaining disks when one dies.
* RAID-6 – similar to RAID-5, but you can tolerate two dead disks at the cost of losing two disks’ worth of capacity. The math is a lot more complicated though.
Generally all disks need to be the same size, or else each disk will be treated as only being as small as the smallest disk in the set. It’s necessary because of how the schemes work.
Amazon’s “RAID servers” are just servers that have additional hardware to do the work of RAID. Otherwise it looks like an normal computer. You can buy such add-on cards but they’re usually expensive, a few hundred dollars to start with and easily $1,000 for the good ones. The RAID hardware provides the illusion to the computer that there is only 1 hard drive, but in fact there are several and all the RAID protection work is done by the hardware. Software versions also exist but must be set up by the user during installation of the operating system and may have other limitations.
Whether RAID makes a system run faster or slower depends on the workload. Reading data is usually faster because RAID controllers now have flexibility in selecting which hard drive to read data from, or data is distributed across multiple disks allowing them to work together to run faster. However writing may be slower because multiple disks need to participate and there is math to be done. If a hard drive has died then RAID-5 and 6 get slower because the lost data must be recalculated and all the other disks need to participate in the recovery of the lost data.
Latest Answers