Old SCSI RAID drives VS new ATA-133 drive
Now that I understand what RAID does, I've been looking through my old discarded hardware in which I found a RAID SCSI controller.
I previously didn't understand it's functionallity, so I through it in my parts box.
The controller is a Adaptec AAA-132/133, and I have three SCSI drives (ST39173W 9.1Gbytes 7,200-RPM internal transfer rate of up to 22.5 Mbyte per second).
Would it be faster using these three old SCSI drives in RAID 0 than using a newer ATA-133 7,200-RPM drive?
I previously tried looking for a driver to get this working on windows XP or Windows 2003, but had no luck finding it.
Does anyone have a method or drivers to get this to work with XP or 2k3?
I'm also looking to see if I can get it to work with Linux (Fedora).
I previously didn't understand it's functionallity, so I through it in my parts box.
The controller is a Adaptec AAA-132/133, and I have three SCSI drives (ST39173W 9.1Gbytes 7,200-RPM internal transfer rate of up to 22.5 Mbyte per second).
Would it be faster using these three old SCSI drives in RAID 0 than using a newer ATA-133 7,200-RPM drive?
I previously tried looking for a driver to get this working on windows XP or Windows 2003, but had no luck finding it.
Does anyone have a method or drivers to get this to work with XP or 2k3?
I'm also looking to see if I can get it to work with Linux (Fedora).
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Found these cached Google pages that suggest that you can use the NT driver from the URL above:
http://72.14.203.104/search?q=cache:rHjD3b5DJbYJ:www2.ntcompatible.com/cdetail11858.html+adaptec+aaa-132+linux&hl=en&ct=clnk&cd=8&client=firefox-a
http://72.14.203.104/search?q=cache:hIdcOyF2D-MJ:www.ntcompatible.com/Adaptec_AAA-132133_c11858.html+adaptec+aaa-132+linux&hl=en&ct=clnk&cd=1&client=firefox-a
So you may find that Linux supports this card natively with one of the ASPI drivers.
http://72.14.203.104/search?q=cache:rHjD3b5DJbYJ:www2.ntcompatible.com/cdetail11858.html+adaptec+aaa-132+linux&hl=en&ct=clnk&cd=8&client=firefox-a
http://72.14.203.104/search?q=cache:hIdcOyF2D-MJ:www.ntcompatible.com/Adaptec_AAA-132133_c11858.html+adaptec+aaa-132+linux&hl=en&ct=clnk&cd=1&client=firefox-a
So you may find that Linux supports this card natively with one of the ASPI drivers.
Finally, here's a blurb about the management software:
http://www.adaptec.com/worldwide/support/supporteditorial.jsp?sess=no&prodkey=array_notes&language=English+US
http://www.adaptec.com/worldwide/support/supporteditorial.jsp?sess=no&prodkey=array_notes&language=English+US
I wouldnt hesitate to use a new ATA133 than the old SCSI drives. A single good mirror would be more reliable on two ATA drives than older SCSI, plus you get 10-30X more disk space. That is my bias, which comes from much hands on experience with SCSI raids, and their manifold problems, not from reading "specs" that hype things up, inflated performance that doesnt come out in the real world, i.e. experience.
I have to disagree with you - SCSI outperforms ATA. Period. And a SCSI RAID 0 set (although very dodgy as it has no redundancy) will massively outperform a single ATA disc. A RAID 5 set will outperform the ATA disc by a factor of between 2 and 3 times on reads, and will probably either be about the same perfromance as the single ATA disc or slightly slower.
Just a note on ATA133. The '133' denotes the bus speed. 133Mhz was a Maxtor thing, and not all IDE controllers support it. The '133' is also misleading - it is the speed that data is transferred to or from the controller on the disc. The data transfer rates on and off the disc are much, much slower and has more to do with rotational speed and the disc geometry than the external bus speed -which has nothing to do with the internal transfer rates. 10-20MB/sec would be typical depending on the age and model and of the drive
This link is the data sheet for an ATA133 drive: http://www.maxtor.com/portal/site/Maxtor/menuitem.ba88f6d7cf664718376049b291346068/?channelpath=/en_us/Products/Desktop%20Storage/DiamondMax%20Family/DiamondMax%20Plus%208
Note that they don't quote an internal data transfer rate - which is the one that matters...
Just a note on ATA133. The '133' denotes the bus speed. 133Mhz was a Maxtor thing, and not all IDE controllers support it. The '133' is also misleading - it is the speed that data is transferred to or from the controller on the disc. The data transfer rates on and off the disc are much, much slower and has more to do with rotational speed and the disc geometry than the external bus speed -which has nothing to do with the internal transfer rates. 10-20MB/sec would be typical depending on the age and model and of the drive
This link is the data sheet for an ATA133 drive: http://www.maxtor.com/portal/site/Maxtor/menuitem.ba88f6d7cf664718376049b291346068/?channelpath=/en_us/Products/Desktop%20Storage/DiamondMax%20Family/DiamondMax%20Plus%208
Note that they don't quote an internal data transfer rate - which is the one that matters...
ASKER
>> I wouldnt hesitate to use a new ATA133 than the old SCSI drives. A single good mirror would be more reliable on two ATA drives than
>>older SCSI
For this paticular requirement, I'm looking more for speed than I am for reliability.
>>That is my bias, which comes from much hands on experience with SCSI raids, and their manifold problems, not from reading "specs" that
>>hype things up, inflated performance that doesnt come out in the real world, i.e. experience.
Are you saying from your experience, that you don't see the speed gain by using SCSI raids?
>>older SCSI
For this paticular requirement, I'm looking more for speed than I am for reliability.
>>That is my bias, which comes from much hands on experience with SCSI raids, and their manifold problems, not from reading "specs" that
>>hype things up, inflated performance that doesnt come out in the real world, i.e. experience.
Are you saying from your experience, that you don't see the speed gain by using SCSI raids?
ASKER
>>I have to disagree with you - SCSI outperforms ATA. Period. And a SCSI RAID 0 set (although very dodgy as it has no redundancy) will
>>massively outperform a single ATA disc
Are there any good free programs to test HD performance?
>>massively outperform a single ATA disc
Are there any good free programs to test HD performance?
>Are there any good free programs to test HD performance?
This is good: http://www.iometer.org/
>For this paticular requirement, I'm looking more for speed than I am for reliability.
If that's the case, then use RAID 0 - but beware! Your hardware is old, so you have a higher chance of losing a disc and therefore all your data...
This is good: http://www.iometer.org/
>For this paticular requirement, I'm looking more for speed than I am for reliability.
If that's the case, then use RAID 0 - but beware! Your hardware is old, so you have a higher chance of losing a disc and therefore all your data...
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
rindi & meyersd,
Good points, however I just plan to use the drives to do builds and run Doxygen on a major projects. They will not be holding important data, but I want something with really fast read and write access.
http://www.pcguide.com/ref/hdd/perf/raid/levels/singleLevel5-c.html
I notice the above link states the following:
**************************
Random Read Performance: Very good to excellent; generally better for larger stripe sizes. Can be better than RAID 0 since the data is distributed over one additional drive, and the parity information is not required during normal reads.
**************************
Doesn't this assume that RAID 0 is only using two drives?
Can't you use RAID 0 with three drives, and if so, wouldn't it be faster than RAID 5 for reading?
Good points, however I just plan to use the drives to do builds and run Doxygen on a major projects. They will not be holding important data, but I want something with really fast read and write access.
http://www.pcguide.com/ref/hdd/perf/raid/levels/singleLevel5-c.html
I notice the above link states the following:
**************************
Random Read Performance: Very good to excellent; generally better for larger stripe sizes. Can be better than RAID 0 since the data is distributed over one additional drive, and the parity information is not required during normal reads.
**************************
Doesn't this assume that RAID 0 is only using two drives?
Can't you use RAID 0 with three drives, and if so, wouldn't it be faster than RAID 5 for reading?
>however I just plan to use the drives to do builds
Go with RAID 0 if your main requirement is performance. Just make sure you have an image or good backups. :-)
With RAID 0, you stripe the data over the three drives. If you configure RAID 5, you stripe data plus parity over the three drives - so you end up with two drives used for data and one used for the parity information.
PC Guide has got that bit wrong. RAID 0 outperforms RAID 5 at the expense of data integrity. So, yes you can use three drives in your RAID 0 set.
Go with RAID 0 if your main requirement is performance. Just make sure you have an image or good backups. :-)
With RAID 0, you stripe the data over the three drives. If you configure RAID 5, you stripe data plus parity over the three drives - so you end up with two drives used for data and one used for the parity information.
PC Guide has got that bit wrong. RAID 0 outperforms RAID 5 at the expense of data integrity. So, yes you can use three drives in your RAID 0 set.
Glad to be of assistance!
ASKER
Thanks
your welcome
Some usefull information:
http://zones.advisor.com/doc/05343