Awesome review! I have been hearing lots of praise from folks that bought an Areca controller and now I can see why. I am getting an ARC-1230 with multilane SATA soon, so this review is just want I needed!

All that performace data is great but I would have liked to hear more about the drivers/utilities/BIOS for that card. I think you could have also concentrated a bit more on the not so commonly seen features, it's not like there are that many SATA controllers that support RAID6!

Next time please also test the fault tolerance and how long an array rebuild takes. This controller can have the best performace in the world but if it craps out during a rebuild it's of no use.

I own this card. its awesome
very low CPU overhead with great responsiveness and a web based browser that doesnt need to be restarted very often.

And again, going to post this !. Have you people ever worked on a server system or a workstation, doesn't look like it

I am sick and tired of reviewers who send 16 pages of an article on A RAID controller talking about performance and not once testing the controllers claim to Fault Tolerance. In case the reviewer missed it, RAID stand for REDUNDANT array of inexpensive disks, so the majority of people using these kind of controllers will be doing so for the fault tolerance aspect. How about dedicating a couple pages to single disk failure and multi disk failures and recoveries instead of almost meaningless performance graphs ?Working with these controllers daily, I know alot of the redundacy aspect are very overrated.

To test the limits of an "8-channel" RAID card you need 8 drives. 4 drives is OK in itself (small arrays, home use), but I'm sure the card could have done much better with 8.
8x Raptor150's would have STR in the range of 340->680.
What is the " ... lack of 'optimizations' are the constituents of the success!"
The RAID1 & RAID10 STR scores showed no optimisations and reduced performance from the potential (RAID1 read STR should be similar to RAID0 read STR if optimised for speed). If you read both and compare RAID1 data then it would reduce the potential but increase integrity ?
RAID10(4HD) did have some advantage over RAID0(2HD) in a few other tests (FC).
No simple tests regarding additional features (hot-plug, rebuild, capacity expansion etc.). No simple test regarding cache size/use(onboard RAM). No simple tests of restricting interface, RAM speed or controller CPU speed. No indication if driver still loads the system CPU.
These can be done elsewhere (some using similar/previous HW), referenced and mentioned in passing/conclusion.

young, yet very talented, Areca.

Young? Areca is 15 years on the market.

Raid6 is raid5 with two copies?
Oh, go and buy a book...
Raid6 stores date + P + Q, where P != Q, a raid6 reonstruct requires divisions and tends to be slow. When testing raid6, test the reconstruct as well!

This card sux. I'm using Areca 24 port card, gave me nothing but trouble. Had to rebuild the 14TB array 5 times.

While their cards do seem to perform great, Areca's customer service leaves a lot to be desired.