Port multiplier performance - external

[nhtshot]

I just wanted to post this for any of you considering using something like this:

http://www.newegg.com/Product/Product.aspx?Item=N82E16817332017

I recently wanted to add more space to my home-server and was out of space in the case. So, I bought an external enclosure with a sata port-multiplier. My logic was that since a single sata link is 3.0Gbps, it should be fast enough to run 5 drives with out much performance loss.

In the end, I've come to find out that the controller used makes a very big difference in the resulting performance.

My server M/B has a built-in eSata port, attached to a J-micron chip. Originally, I plugged the array into here and it seemed to work fine. However, the performance was quite disappointing. ~40MB/sec according to Hdparm.

I debated what to do with it, since I'd already bought the enclosure and the drives. I decided to try it on a different card. I threw in a spare Sil-3132 based PCI-Ex1 eSata controller.

Wow, the difference was amazing.

Now, it does ~160MB/sec.

Moral of all this:
Don't discount eSata enclosures with port multipliers due to performance concerns. They're actually pretty speedy. Furthermore, don't assume that your M/B's eSata port is anything other then junk.

Has anyone else had bad experiences with the jmicron contollers?

 

[jtg1993]

Your onboard eSata might of been 1.5GBps, that would explain the crappy speeds.

 

[epimetheus]

OP, have you looked at the performance when multiple drives are being accessed at the same time thru the port multiplier? I'm curious what the performance is like copying from one drive on the port multiplier to another on the port multiplier. This is where the speed compromises of a port multiplier are going to be seen AFAIK. Granted, this wont make a difference depending on how you use it. If you're using it for WHS, you would probably rarely get in a situation where any compromise would be apparent to the end user.