Raid tutorial

Status
Not open for further replies.
D

DelJo63

RAID: Redundant Array of Inexpensive Disks

Classic hardware is a SCSI controller with multiple Scsi HDs. Today, there are cheap
IDE implementations, but the raid functionality is in software on the host -- impacting
both reliability and performance. The latest technology is FDDI fiber channel.

Hot Swapping: the ability to extract a failed HD and to replace it with another without
disconnecting power or the connection to the system. (Not possible in the IDE implementation).

RAID comes in several variations, 0, 1, 2, 3, 4, 5, 6, 10, 50, 0+1, and jbod

JBOD:
Concatenation (JBOD or SPAN)

JBOD stands for Just a Bunch of Drives. The controller treats each drive as a stand-alone disk, therefore each drive is an independent logical drive. JBOD does not provide data redundancy.

Concatenation or Spanning of disks is not one of the numbered RAID levels, but it is a popular method for combining multiple physical disk drives into a single virtual disk. It provides no data redundancy. As the name implies, disks are merely concatenated together, end to beginning, so they appear to be a single large disk. This mode is sometimes called JBOD, or "Just a Bunch Of Disks".

see full descriptions and illustrations for each of these.

see also Wiki


[edit]Raid-0:
  • As soon as some see hi performance, some users will immediate jump for joy and dive headlong into the
    deep end of the pool without considering 'What If ...'. Over and over again I've seen systems where
    an HD error on raid-0 causes a TOTAL LOSS and the necessity to rebuild. If there's no backup,
    then all data is gone! Even Microsoft does not recommend a RAID on the boot volume.

    What is raid-0 hi performance anyway?
    HD performance (ignoring an onboard cache) is comprised of three factors:
    • time to move the disk arm to the correct location (called seek time)
    • time to let the disk rotate to the correct record (called latency)
    • time to read the data into memory (transfer)
    These timings are (relatively) 100, 10, 1 respectively;
    ie: it takes 10 times longer to move the arm than to locate the record.

    Raid-0 allows multiple seeks to be started at the same time. Without raid, the time required would be
    time for X records = X * (seek+latency+transfer)​
    With a raid-0, the time would be
    time for X records = (X *seek)/D + (X * (latency+transfer))
    where D is the number of HDs in the raid. Clearly the more HDs the better the performance.​
    Cost / Benefit
    The cost increase is linear for adding HDs but multiplicative when the backup media requirements are considered.
    Then too it takes major time to backup a logical volume of X 500gbHDs in a raid-0 configuration.
    The performance improvement is fractional unless there are a large number (ie millions) of sustained I/Os
    (as in a server like Google which would be far more likely to implement raid-0+1).

    Recommendation for home users: Forget it. Just use 10k rpm, caching HDs and make your life simple.
[/edit]
 
Status
Not open for further replies.
Back