On a MS Windows machine IDE Controller & Disks it is very slow.
Compared to a SCSI Raid on a Linux box, it is extremely slow.
Linux software RAID (md) driver has been benchmarked in RAID 1 and RAID 0 configs to be faster than similar configured IDE RAID controllers, RAID 5 speed would be totally dependent on the situation. I'm sure his MP3 and TV video collection is 95% read access, so RAID 5 in software would be fine once the initial build and copy finishes.
Best performance is achieved when data is striped across multiple controllers
with only one drive per controller
Even with IDE a single controller there's 2 interfaces, so you can have 2 drives without them blocking each other. With SCSI each drive disconnects from the chain while it does the transaction and each device can have multiple transactions queued up, so essentially they don't get in each other's way even if they're on the same wire. And a basic SCSI3 controller can handle 15 devices per interface.
Twice the Read transaction rate of single disks, same Write transaction rate as
single disks
Depends on the implementation, sometimes the write time is twice as slow because it waits for both to write to ensure data integrity. But with all the levels of caching on a computer it won't really be twice as slow.
Under certain circumstances, RAID 1 can sustain multiple simultaneous drive failures
I've not seen a setup that lets you have a single drive with 2 mirrors, although I don't see technically why it can't be done.
Typically the RAID function is done by system software, loading the CPU/Server and
possibly degrading throughput at high activity levels. Hardware raid always is
strongly recommended
There's no real overhead to load on the system, it issues 2 writes per memory page instead of 1, you're still stuck waiting on disk I/O.
May not support hot swap of failed disk when implemented in "software"
A lot of hardware solutions don't support hotswap, especially IDE ones.
Highest Read data transaction rate (RAID 5)
Depends. If I have 16 drives in RAID 0 and 16 drives in RAID 5 I bet the RAID 0 set will be faster, all else equal.
Disk failure has a medium impact on throughput
Unless you have a spare disk allocated a disk failure means the array is down until you get a replacement, unlike RAID 1.
Difficult to rebuild in the event of a disk failure (as compared to RAID level 1)
Should be just as easy as the others, add the new disk and let it rebuild the array.
Individual block data transfer rate same as single disk
The individual block transfer rates are the same across the board, you only have a single block on any disk. Unless you're talking about a different measurement than the 'block' I've always used to mean a physical disk sector, i.e. 512 bytes (except on really odd hardware). If you mean filesystem cluster or just 'chunk of data' then yes you're right.