• Embed Doc
  • Readcast
  • Collections
  • 1
    CommentGo Back
Download
 
New Page 1
Recently one of my friend asked me what do I think of Raid 0 which he is thinking of implementing in hisserver / workstation. So I grabbed a few books and went through and this is what I found.
 
"RAID" stands for "
Redundant Array of Inexpensive Disks
", (now being reinvented the term todescribe a
redundant array ofindependentdisks
as a means of disassociating a "low cost"expectation from RAID technology) and refers to several different ways to use multiple hard drives toeither increase performance, reliability or both. RAID is typically transparent to you or the operatingsystem, other than the fact that there is a dedicated piece of hardware, known as a RAID controller, thatmanages the distribution of data across the drives.
 
 Various standard configurations of RAID drives, typically referred to as "RAID arrays", have been labeledwith numbers to differentiate the differing arrangements.RAID 0 uses two (or more) drives to increase performance.
 
The approach is this: when you write data to the logical (say, "C:") drive, the data is interweaved insome way such that half of it is written to one physical drive, and half on the other. Effectively, thisdoubles the speed of your hard disk, since the computer can essentially be writing to both at the sametime, instead of writing the same amount of data in two steps to a single drive.
 
Sounds like a nifty approach, right? You get the same amount of space as you would with two individualdrives, but you get it twice as fast.
 
The problem is that you get twice as much of something else as well: chance of failure and catastrophicdata loss.
 
If one drive has, say, a 1 in 1,000 chance of failure, then two drives have a 2 in 1,000 chance.
 
Now, if you have two separate drives you still have that same 2 in 1,000 chance ofoneof them failing -but all you lose is what's on that one drive that fails. In a RAID 0 configuration, you've doubled yourchances of failure, and when failure happens, all of your dataon both drivesis lost.
 
Think of it this way: if we take a simple sentence like:
 
RAID 0 configurations are a bad idea.
 
and sliced it in half such that every other letter was "on alternate drives", this might be what appears oneach drive:
 
Drive 1: R I   o f g r t o s a e a b d i e .Drive 2: A D 0 c n i u a i n  r   a  d a
 
Together they're easy (and quick!) to combine into the original sentence. Take away either one and theother is so much gibberish.
 
RAID 0 is valuable as a speed enhancement only if you're willing to take double the risk of failure. Thatmeans either that it's OK if the failure happens and you lose the data, or you're conscientious aboutmaking sure that it's backed up regularly and thoroughly.
 
What's frustrating, in part, is that once you have a RAID controller anyway, with the addition of a third(inexpensive) drive you can get all that speed as well as fault tolerance using different RAIDconfiguration.
 
Using RAID 5, a third drive contains "parity" information - kind of the "C" in an "A + B = C" type of equation - lose any of A, B or C and it can be recalculated from the other two. In the world of hard disk 
Page 1
 
New Page 1
drives, that means a drive can fail completely and your system continues to operate while the disk isreplaced. RAID controllers often include "hot swap" ability, so that could even be literal: the drive couldbe replaced without stopping or restarting the system.
 
That's just one example - RAID as a concept actually includes several different arrangements that youcan use to balance performance, fault tolerance and risk.
 
Just make sure that if you're using RAID 0, you know what you're getting into.
 
There are various combinations of these approaches giving different trade-offs of protection against dataloss, capacity, and speed. RAID levels 0, 1, and 5 are the most commonly found, and cover mostrequirements.
 
RAID 0 (striped disks)distributes data across several disks in a way that gives improved speed at anygiven instant. If one disk fails, however, all of the data on the array will be lost, as there is neitherparity nor mirroring.
 
RAID 1mirrors the contents of the disks, making a form of 1:1 ratio real-time backup. The contentsof each disk in the array are identical to that of every other disk in the array.
 
RAID 5(striped disks with parity) combines three or more disks in a way that protects data againstloss of any one disk. The storage capacity of the array is reduced by one disk.
 
RAID 6(striped disks with dual parity) can recover from the loss of two disks.
 
RAID 10 (or 1+0)uses both striping and mirroring. "01" or "0+1" is sometimes distinguished from"10" or "1+0": a striped set of mirrored subsets and a mirrored set of striped subsets are both valid,but distinct, configurations.
 
RAID systems with redundancy continue working without interruption when one (or possibly more,depending on the type of RAID) disks of the array fail, although they are then vulnerable to furtherfailures. When the bad disk is replaced by a new one the array is rebuilt while the system continues tooperate normally. Some systems have to be powered down when removing or adding a drive; otherssupporthot swapping, allowing drives to be replaced without powering down. RAID with hot-swapping isoften used inhigh availabilitysystems, where it is important that the system remains running as much of the time as possible.
 
RAID is not a good alternative tobacking updata. Data may become damaged or destroyed withoutharm to the drive(s) on which they are stored. For example, some of the data may be overwritten by asystem malfunction; a file may be damaged or deleted by user error or malice and not noticed for daysor weeks; and, of course, the entire array is at risk of physical damage.
 
Following is a brief summary of the most commonly used RAID levels.
[3]
Space efficiency is given asamount of storage space available in an array of n disks, in multiples of the capacity of a single drive.For example if an array holds n=5 drives of 250GB and efficiency is n-1 then available space is 4 times250GB or roughly 1TB.
 
Level
 
Description
 
Minimum# of disks
 
SpaceEfficiency
 
Image
 
"
Stripedset withoutparity
" or"
Striping
". Provides improved performanceand additional storage but no redundancy orfault tolerance. Any disk failure destroys thearray, which has greater consequences withmore disks in the array (at a minimum,catastrophic data loss is twice as severe
Page 2
 
New Page 1
RAID0
 
compared to single drives without RAID). A single disk failure destroys the entire arraybecause when data is written to a RAID 0drive, the data is broken into fragments. Thenumber of fragments is dictated by thenumber of disks in the array. The fragmentsare written to their respective diskssimultaneously on the same sector. Thisallows smaller sections of the entire chunk of data to be read off the drive in parallel,increasing bandwidth. RAID 0 does notimplement error checking so any error isunrecoverable. More disks in the arraymeans higher bandwidth, but greater risk of data loss.
 
2
 
n
 
RAID1
 
'
Mirroredset without parity'
or'
Mirroring'
. Provides fault tolerance fromdisk errors and failure of all but one of thedrives. Increased read performance occurswhen using amulti-threadedoperatingsystem that supports split seeks, as well as avery small performance reduction whenwriting. Array continues to operate so longas at least one drive is functioning. UsingRAID 1 with a separate controller for eachdisk is sometimes calledduplexing.
 
2
 
1 (size of thesmallestdisk)
 
RAID2
 
Hamming codeparity
. Disks aresynchronized and striped in very smallstripes, often in single bytes/words.Hamming codeserror correctioniscalculated across corresponding bits ondisks, and is stored on multiple parity disks.
 
3
 
RAID3
 
Striped set with dedicated parity
or
bitinterleaved parity
or
byte level parity
.
 
This mechanism provides fault tolerancesimilar to RAID 5. However, because thestrip across the disks is a lot smaller than afilesystem block, reads and writes to thearray perform like a single drive with a highlinear write performance. For this to work properly, the drives must have synchronisedrotation. If one drive fails, the performancedoesn't change.
 
3
 
n-1
 
RAID4
 
Block level parity.
Identical to RAID 3, butdoes block-level striping instead of byte-level striping. In this setup, files can bedistributed between multiple disks. Each disk operates independently which allows I/Orequests to be performed in parallel, thoughdata transfer speeds can suffer due to thetype of parity. The error detection isachieved through dedicated parity and isstored in a separate, single disk unit.
 
3
 
n-1
 
Page 3
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...

Thank You Boss.I'm also a student of Mr.Hati..

You must be to leave a comment.
Submit
Characters: ...