http://www.cheetahdrives.com/

[Dïscfärm]

http://www.cheetahdrives.com/

 

[P5-133XL]

Gimmie a break - Short stroking the 10K drives to increase performance and decrease reliability. They are using a 14 (10K)/11 (15K) drive raid-0 array and thus are playing games with transfer rates too; With that many drives the transfer rate will be limited by the PCI bus.

Seagate should not need to be playing such games with the statistics. This is just junk!

 

[blakerwry]

That's the way I felt about it at first too... but I have herad of people short stroking 10k drives in a real wolrd situation (not just testing). So I figure this may have some basis in reality.

The only thing you can conclude from this is that you may be able to get similar level of performance while getting better reliability and cost savings (along with rack space savings) by using 15k drives instead of short stroked 10k drives.

While that is a lot of data, no doubt some machines can handle it.

 

[Howell]

**Implemented using the common practice of storing data on the outer diameter of the media (destroking) to enhance the performance of the 10K drives. In this configuration, the 10K RPM was destroked to 78.6% of its full capacity.

 

[time]

The fact that they've almost certainly short-stroked the 15K drives as part of the design, doesn't seem to trouble them. :roll:

It's a bit like arguing that 80GB drives are twice as reliable as 40GB, because you only need to use half as many ...

I agree with P5-133XL.

 

[blakerwry]

The fact that they've almost certainly short-stroked the 15K drives as part of the design, doesn't seem to trouble them. :roll:

It's a bit like arguing that 80GB drives are twice as reliable as 40GB, because you only need to use half as many ...

I agree with P5-133XL.

If you have a terrabyte array then yes, I'd say that the 80GB array would prove to have less failures (assuming you're comparing 80gb vs 40gb drives of the same model)