Tom's has got the lowdown on Enterprise storage offerings, including the Barracuda ES series tailored for 24/7 operation, and the new WD Raptor 74B ADFD. Should you open a SATA slot in your tower for them?
SCSI is the protocol that a drive interface can use. It actually has little impact on drive reliability. In fact, I bet if you looked at the numbers, you would find that modern SCSI drives have about the same failure rate as SATA drives from the same manufacturer. The difference is, that when SCSI arrays are properly implemented, you have little to no chance of losing data when an individual drive fails, and thus, 24/7 operation becomes a viable operation, as the loss of a single member (and sometimes, even multiple members) of an array will not impact immediate access to data, and can be replaced and restored.
As we start to see more and more SATA raid solutions (along with SAS in the wings, which can use SATA drives as well as SAS drives), I think we will start to see more multi-drive arrays used in desktop systems, let alone entry level workstations and servers. Most enthusiast motherboards already include level 0 and level 1 RAID options using SATA drives. Some even offer options to build a level 5 arrray (granted, most of these use a software RAID which is a PITA to reconstruct if you should happen to lose a member).
While SCSI is reliable, SCSI isn't affordable for everyone. Sure, most businesses will stay with SCSI or SAS. They can afford the high prices of specialized hardware and drives. Most of the rest of us are more than happy to see some of these redundant array solutions finally trickling down to hardware we can actually afford. And I wouldn't be surprised to find one or two data centers willing to try out a cheaper solution for their less mission critical storage needs, where availability on a budget is more important than access speed. So it makes sense that Seagate should start offering solutions for this market.
1. Failure to mentioned how Barracuda ES differs from 7200.10 through firmware optimisation, leaving only educated readers to figure it out in IOMeter tests. (What the whole article should've been based on)
Jason Zushman and Achim Roos, if you're gonna play with the enterprise boys, at least give them something worthwhile to read in the conclusion section and not just raw data.
i'm surprised at how much faster the New Revision raptor is than the old 8mb Cache model. I might just feel a quick upgrade coming up.
A few loses here and few gains(read STR) there. Meh, next to no difference at all in either enterprise or single-user usage of real life scenarios.
You're just buying for the extra space in WD1500AD and possibly better GB/$ with WD740ADFD.
To me, this review is like comparing cache on a processor to RAM.
And yes...most servers...SCSI. If you want performance that is the way to go. 15k/rpm
Go ask fiber storage companies what they use. ATA for backup arrays yes, but SCSI is the path of speed.
This seemed a somewhat silly and obvious article to me.
Benchmarking these two drives together is like racing a truck against a ferrari, for one thing. In my position, I would rather be pointing my collegues and managers at a benchmark between an array of Raptors vs a fiberchannel array of 10,000rpm FC drives - same number of disks each. *Those* numbers would be interesting. Does a $5000 array of raptors perform the same as a $30,000 SAN? That would be news for "The Enterprise"
A 7,200rpm FATA array vs a few of the seagates raided together would also be worth reading.
For another, this article doesn't tell us anything new. What's the difference between a Raptor ADFD and the previous Raptor models? Is it the cache size? The interface? The article doesn't tell us, except to say that the transparent faceplate model has been dropped. What's new and exciting about the new barracuda, that hasn't been said in other articles?
I still see Tom's as the world leader in hardware reviews. But that lead wont be maintained if you keep cranking out trivial articles that don't say anything I can't read from the product brochure.
So you've got 16M cache on your drive and you say you don't see any difference between SATA 300 and 150???
WD needs to get on board and bring all of that claimed high performance to the interface (for those little read and write burst that we sometimes must do). We might see some real-life performance improvements then, unless there are other issues that keep them from upping the transfer rate (most other manufacturers don't seem to have the same issues/problems with the higher rates)...
Forgot to do this one though 
