Why is RAID 0 classed as RAID when it's not redundant?

I've worked in IT quite a number of years, so I know what a RAID array is, what RAID 0 is, RAID 1, 5, 6, 10, 50, 60, etc., but something sprung to mind in a recent conversation at work; if RAID stands for redundant array of independent (or inexpensive) disks, then why is RAID 0 classed as RAID at all and not just a striped array?

Having data striped across multiple disks on the one array offers no redundancy whatsoever so why is it classed as a RAID array? Surely the lowest number should be RAID 1 (mirrored) as that's when redundancy actually starts?

• The 0 doesn't indicate a raid level, it indicates striping. Same for 0,10,50,60, etc. Dec 17, 2015 at 13:38
• I think it depends on what definition you abide. dictionary.reference.com/browse/redundant mentions: 2. being in excess; exceeding what is usual or natural. - So if you accept this definition then RAID-0 is legal terminology because the extra disk is not needed for normal operation. Dec 17, 2015 at 14:57
• Same reason zero pounds is a sum of money, but it's not worth anything. Dec 18, 2015 at 23:18

The lowest form of RAID is RAID 1. RAID 0 was added well after RAID was defined (can't find reference to a date for this though)

The 0 in RAID 0 is used to signify that actually it isn't considered redundant. Think of it as more True/False where 0 is False.

• Seems so, reading Wikipedia has dates for RAID-4 (the first apparently), RAID-1, RAID-5, RAID-2, shame it doesn't mention when RAID-0 came along, maybe it should be called NRAID0 - Non-Redundant ;) Dec 17, 2015 at 10:43
• The original 1988 paper from Patterson, Gibson, Katz that describes RAID-1 through RAID-5 is here: cs.cmu.edu/~garth/RAIDpaper/Patterson88.pdf . A 1994 paper that adds references to RAID-0 and RAID-6 is here: web.eecs.umich.edu/~pmchen/papers/chen94_1.pdf Dec 17, 2015 at 14:46
• So if RAID-0 means RAID is false, why is it still called RAID?
– Zack
Dec 17, 2015 at 16:34
• @Zack - because the rest of the acronym fits (array of inexpensive disks) and the use of acronyms frequently grow beyond their designed intent when it is appropriate. Take HTTP, it is in the same boat, as it is rarely used to transport hypertext, but also json, video, audio, graphics... nobody would go "zomg!!! that's not hypertext I need find a different protocol." Dec 17, 2015 at 20:53
• Sooo.... the writers of the RAID spec were Python developers? Dec 18, 2015 at 15:27

RAID is just a name with a meaning that changed over time.

The important part is that the underlying technology and mechanisms are the same for the RAID levels, so you use the same controller (or piece of software, e.g. mdraid) to achieve all RAID levels.

• I guess the R in RAID has always been the word Redundant though? Indeed, a RAID controller can do striped, mirrored, etc. even though striped is not redundant so makes sense. Dec 17, 2015 at 10:40
• Yes, but really, it's just a name. I stopped caring for names decades ago because of all the useless marketing we tend to get thrown in...
– Sven
Dec 17, 2015 at 10:44
• @GavinBurke RAID is a thing that adds redundancy to disks. RAID 0 is a variant of RAID that happens to not actually be redundant, but it's otherwise still RAID, so they called it RAID 0. Dec 17, 2015 at 21:25
• @GavinBurke In modded Minecraft there's an item called an "unstable ingot", which violently explodes shortly after you craft it. Later on, they added a way to make ones that don't explode... they're identical to unstable ingots, except they don't explode, so they're called "stable unstable ingots". Same kind of idea in action. Dec 17, 2015 at 21:26

As others have suggested, RAID 0 could be taken as level 0 meaning zero redundancy. It is referred to as RAID even though there is no redundancy for two other reasons:

• It is usually defined and talked about in the same contexts, so the name stuck. The same can't be said for JBOD, but such arrangements don't tend to get described along with RAID levels as they have less in common (RAID 0 at least involves striping as found in the likes of RAID 5).

• While RAID 0 offers no redundancy on its own, it is often used as part of a composite RAID arrangement that does offer redundancy. RAID 10 is the most common, combining 0 for speed with 1 for redundancy. RAID 50 and RAID 60 are less common, but equally valid except where the potential write performance hit (due to 5/6's read-checksum-write pattern) is a concern.

Normaly you talk about a RAID Level. So i you say a System is RAID-0 it does not mean it is 'redundant with type 0', it does only mean it is of 'RAID Level 0', witch means no raid at all.

Because "not redundant" is a valid point on the scale of how redundant something is.

Pure marketing.

<reminisce> I remember when file servers for local area networks started appearing on the market. And yes, RAID 0 was one of the possible configurations. RAID was such a hefty buzzword at the time that managers would just simply ask "well does the server have RAID?" without knowing exactly what that meant. So the crafty salespeople started relabelling striping as "RAID 0" and thus they could "truthfully" say "why yes it does!" </reminisce>

RAID... it was a punchy acronym. Not like, say SCSI, (pronounced "scuzzy.") It was one of those few acronyms that an IT manager could fling out at a high level budget meeting to dazzle others into approving a purchase. You want to get your slice of the budget pie? You've got to have good acronyms. Just try saying "we've just got to get SCSI" in a meeting like that. See how far you get.