Talk:RAID/Archive 7

Archive 1

←

Archive 5

Archive 6

Archive 7

"Inexpensive" vs "Independent"

It correctly states in the article that RAID stands for "Redundant Array of Inexpensive Disks" but then the introduction of the article uses "Independent" instead of "Inexpensive". "Independent" doesn't even make sense. There's nothing independent about a drive in RAID 0, for example. One goes and the data goes. "Inexpensive" is more correct since it describes what is being done. You are taking two or more disk drives and creating a faster, more reliable drive for a lower cost than an individual drive of the same type with the equivalent capacity. If we're going to redefine terms for modernization, RAID should stand for "Redundant Array of Inexpensive Drives" now that we live in a world with SSDs which aren't disks.

The fact that manufacturers incorrectly redefined it is interesting, but that's probably more because some marketing idiot got it wrong and just ran with it without even thinking if it even made sense. If including the incorrect, industry term is important, it should be used as an "or" in the overview with "Inexpensive" being in the primary description since that's what the inventors called it, and not the manufacturers misnamed it.

The "inexpensive" originates from the historical distinction between a high-performance, highly reliable – and very expensive – hard disk and a standard, commodity one. Since RAID rendered the reliable drive systems obsolete for a fraction of the cost they've vanished in the 1980s. The "independent" keeps the acronym alive and on the other hand indicates that the commodity drives are independent entities or devices, making them easily field-replacable and commonly even hot-swap. This wouldn't be possible with integrated drives (that some reliable drives consisted of). --Zac67 (talk) 04:42, 30 August 2018 (UTC)

The SLEDs vanished in the 1990s. Tom94022 (talk) 06:23, 30 August 2018 (UTC)

Yes – thanks, SLED was what slipped my mind! --Zac67 (talk) 17:38, 30 August 2018 (UTC)

All the references to the first paragraph say "inexpensive" (including the OED ref I just added), so it would appear that inexpensive should appear first, unless reference for "independent" is supplied in first para. Pol098 (talk) 16:37, 18 April 2019 (UTC)

Ditto

I agree with your point of view. Should the article be updated? vicsar (talk) 23:10, 21 September 2017 (UTC)

It was originally called RAID by the inventors and nowadays it still is - I'll rename it back Jlascar 10:58, 12 November 2017 (UTC) — Preceding unsigned comment added by Jlascar (talk • contribs)

Ditto

Plus one for inexpensive. The introduction of the term RAID was intended to be opposite of SLED (Single large expensive disk.) Unsure where this term "Independent" came from as it does not even apply. If you start taking enough drives out of any RAID setup, you will eventually realize that they are not so "independent." — Preceding unsigned comment added by 2620:15C:203:0:C6D1:A4F0:4D80:129A (talk) 18:17, 30 April 2018 (UTC)

It probably should be both

Although the original usage was "Inexpensive" much of the industry including its consortium, the RAID Advisory Board, adopted "Independent" so I am changing the lede to use both as in BytePyle Tom94022 (talk) 21:36, 15 November 2017 (UTC)

Maybe not

Sometimes Industry is wrong. — Preceding unsigned comment added by 2620:15C:203:0:C6D1:A4F0:4D80:129A (talk) 18:18, 30 April 2018 (UTC)

Indeed, quite often industry gets it wrong: RJ-45 is not an 8P8C modular connector used for Ethernet and a 9-pin D-shell connector is not designated DB-9, to name but two common misuses. 97.86.156.248 (talk) 23:56, 29 August 2018 (UTC)

It should be both as it is

In this case an industry committee redefined the term to reflect the fact that the early RAID subsystems used expensive SCSI HDDs rather than the inexpensive (parallel) ATA HDDs. Somewhat later inexpensive PATA and SATA drives came to be used in "nearline" RAID subsystems while "enterprise" RAID systems continued to use expensive SCSI and SAS HDDs. Today "enterprise" RAID subsystems are almost all expensive SSDs so today "independent" is still the best term not withstanding the term of the original RAID paper. They probably should have changed "disk" to "drive." Tom94022 (talk) 06:23, 30 August 2018 (UTC)

"Inexpensive" is what does not make any sense now or 30 years ago. Do you any of you people contributing to this wiki page have any understanding of how these "independent" disks are used to form an Array of disks? When RAID was first introduced it was primarily used with SCSI drives, which are not what anyone who knows about the technology would characterize as "inexpensive". They were and still are relatively expensive drives as compared to the more common desktop SATA interface. I think what has happened here is some academics have twisted things around because they do not understand the underlying technology and they found something in a document to support their "inexpensive" case. — Preceding unsigned comment added by 63.250.109.122 (talk) 03:48, 20 October 2018 (UTC)

They are not "Independent." You need all of them in order for RAID to be fully operational. It is the Opposite of "Independent." You can call them "Interdependent" if that makes you more happy. — Preceding unsigned comment added by 2620:15C:203:0:C6D1:A4F0:4D80:129A (talk) 23:04, 5 December 2018 (UTC)

Independent is correct. Similar to the "I" in ACID transactions (stands for "Isolated" may Jim Gray rest in peace), RAID depends heavily on the atomicity of drive failures, meaning failures must be "independent" of each other. At the time, only one drive failure at a time could be handled, and RAID assumes the drives fail indepently, which leads to redundant power systems and dual-port drives and controllers. — Preceding unsigned comment added by SeaChanger (talk • contribs) 19:45, 19 January 2019 (UTC)