Fast and Secure: A Comparison of Eight RAID Contro

[sjonnie]

This review was a joke basically. Don't know why companies bother sending stuff to THG if this is the treatment they get. Why didn't they show any data for goodness sake? And comparing the 3ware controller with the others is total nonsense. Not only is it a totally difference hardware design (about which they mention nothing) it is actually a RAID controller and not just a modified IDE controller like the others.

If you want a proper review look at <A HREF="http://www.xbitlabs.com/storage/" target="_new"> http://www.xbitlabs.com/storage/ </A>

 

[Ncogneto]

Isn't the pic that is linked to the 7850 actually a 7450 as it only has 4 channels and not the eight?

The cache controller supports level 5 in addition to the usual RAID modes. This makes the most effective use of disk space. The speed penalty is about 10 percent compared with RAID levels 0 and 1, although, in practice, this will hardly be noticeable. The RAID-0 and RAID-5 results are close together at 56,800 and 56,750 Kbytes per second (RAID-0 compared to RAID-5).

Umm, ok, in which discipline? Reads maybe...but writes????? Your right the review (if one can even call it that) was a joke.

It's not what they tell you, its what they don't tell you!<P ID="edit"><FONT SIZE=-1><EM>Edited by ncogneto on 11/12/02 01:13 PM.</EM></FONT></P>

 

[lhgpoobaa]

True True... seemed a little lacking in hard facts. Then again im not an expert in Raid cards.

<b>LHGPooBaa + Evil Hamster Sidekick: Serving Toms Hardware community for 2 years as of the 11th of November</b>

 

[sjonnie]

Yep, the picture is a 7450.

The RAID-0 and RAID-5 results are close together at 56,800 and 56,750 Kbytes per second (RAID-0 compared to RAID-5).

But of course they fail to explain what on earth they are compairing. As far as I could tell they used a two disc RAID0 setup for most of the tests. But for RAID5 you need a minimum of 3 disks. So did they compare a RAID5 array of 3 discs with a RAID0 array of 2 or 3 disks? A RAID0 array of 3 disks should give a sequential write performance of about 80Mb/s so it looks like they used a 2HD RAID0 array. With this card a RAID0 array of 2 disks gives about the same write performance as a RAID5 with 3 disks.

See the review at <A HREF="http://www.xbitlabs.com/storage/3ware-7850/" target="_new">xbitlabs</A> for serious info on the 3ware 7850

 

[Ncogneto]

But for RAID5 you need a minimum of 3 disks.

Not to mention the three disks they reference are not even all the same. not to mention they fail to explain the diference between a software controller and a hardware controller.

It's not what they tell you, its what they don't tell you!

 

[lootman]

My biggest issue is that they put the 3ware card in there. The price alone should be a tip-off that it doesn't belong. A quick look on the 3ware page shows that the 7850 is now called the 7500-8. This is a midrange server card in a desktop card review. They should have put in a 7500-2 or 4 to be fair to the other cards. I would have loved to have seen a review between the 7500-8, the Promise 6000 and the Adaptec 2400A.

 

[sjonnie]

I would have loved to have seen a review between the 7500-8, the Promise 6000 and the Adaptec 2400A.

You can find a review of those controllers <A HREF="http://www.xbitlabs.com/storage/raid-roundup-2/" target="_new"> here</A>. Note they use the 7810 which doesn't have RAID5 fusion technology - in any event, it is the clear winner.