So this is the thread that everybody has been asking me about...
ok here we go... pictures and all... links and all...
We call ourselves techs but do we care about our data? I remember the days of napster and when my 2.5gb maxtor was full and I was invincible... until it crashed. I realised that I was going crazy . I couldn't sleep well; don't laugh cuz if this happened to me prob a few months ago, you would hear about me in the news.
On setting out to build my ultimate machine, I decided that raid would have to be an integral part of the infrastucture of the system. I penciled out what the specs should be then started my research. I was so clueless about the topology that I did reading like crazy (read your literature and do your homework). I am now 99% raid compliant.
So on with the types of raid:
0 or what we call stripe; increases read and write speeds exponentially ie 2 drives gives 2 x regular speeds, 3 drives gives 3 x regular speeds and so on. This technique has what we call no "fault tolerance" ie if the array is broken, the data is lost. With newer technology though, you realise that a hard drive can give us 3 to 4 years pending cooling is added or operating under extremely cool conditions.
1 or what we call mirror; increases read speeds exponentially as well but has a little drop in write speeds, nothing significant to even notice. This technique is the first of fault tolerance methods that is available throughout the raiding topology. Data is written to two separate hard drives which must be of the same size or will be of the same size of the smaller drive if one is bigger than the other. Essentially if one drive fails then windows/controller will only report it but continue to use the other. Once replaced windows/controller will rebuild the array and resync the drives.
JBOD or Just a Bunch Of Disk; otherwise called spanning, this is somewhat of a raid 0 but does not require drives to be the same size. For example a 20 + 80 + 40 would give you one big volume of 140 with no speed improvement and still suffering the same faith of raid 0 if one drive fails.
5 or Raid with parity; This is basically a raid 0 with parity data written to each disk and a seperate disk which would be used to reconstruct any disk out of the array in case one fails. This raid has fault tolerance
0+1 or Mirroring of a Stripe; kinda self explanatory; hence the minimum of this array would have to be 4 drives. This raid has fault tolerance.
Now I have explained the topology and techniques, on to raid in windows.