Recommend to me: External 2-drive hi-perf hdd enclosure

Jaxidian

Platinum Member
Oct 22, 2001
2,230
0
71
twitter.com
Close but no cigar...
I found this at Newegg and it's very close to what I'm looking for but the Firewire is only Firewire 400 and I'd like Firewire 800.

Features that would be nice:
The RAID option is nice but I'm not willing to pay an arm and a leg for it if it's required in my options. I'm planning on putting a couple relatively large (500 or maybe even 750) 7200.10 drives into this and will be running VMs off of it.

More background info in case you have some good suggestions:
Basically I need to run some virtual machines off of an external hard drive. I need this to be as high-perf as possible (and reasonably priced). I originally thought about having one of the two drives be a Raptor 150 and the other one being a 750gb 7200.10 drive but I've been informed that Raptors run too hot to be reasonable external drives so I've backed off from that idea (but would love to hear some stories if anybody else has done this already).
I'd prefer SATA over IDE just for future upgradability even if there isn't a compelling performance reason to prefer that (though please tell me if for some reason you think it's better for performance to go IDE).
I'd prefer FireWire over USB 2.0 because I hear it's best for external drive performance (especially when the CPU is stressed).
I'd really like to have a dual-drive enclosure that shares 1 FireWire port simply for the convenience of not having to pull around 2 external enclosures plus power adapters. I'd like a dual-drive enclosure because I'd like to run my VMs off of one drive and use the second drive for mass storage, installers, backups, etc. I'd rather not stress the same drive when performing actions both in a VM and with file xfers and such plus I'd also like to have the option to toss a VM on the second drive in case I want to run 2 VMs at the same time (won't happen very often but I'm sure it will happen occassionally).
Oh, and eSATA really isn't an option as none of my laptops have that and that's what this is going to be hooked up to (find me a good/reasonably-priced PC Card that has eSATA and I may change my mind).

Help appreciated!

-Shane
 

Old Hippie

Diamond Member
Oct 8, 2005
6,361
1
0
How about something like this? If that doesn't suit your fancy, just look around the site. It gets kinda confusing because they make just about every configuration (internal and external connections) under the sun. Read the descriptions carefully cause some units come with no trays, no cables, plastic trays.....you get the idea.
 

Paperdoc

Platinum Member
Aug 17, 2006
2,381
310
126
Even if the enclosure has IEEE 1394b (Firewire800), does your computer support that? Or, are you looking at future hardware changes?

I saw some performance benchmarks recently for an external enclosure with only one HDD. The actual measured data transfer rate was almost identical for USB2 and Firewire 400 connections. The rate for eSATA was almost twice as fast. Just FYI. As I read notes, the major difference between USB2 and Firewire connections is that USB2 may be subject to more short-term interruptions in data flow. For things like streaming video this matters, but for data transfers to a HDD I would guess it does not, as long as the overall transfer rate is comparable.

I see you don't have the eSATa option now, anyway. But I also wonder: would any eSATA enclosure share ONE eSATA connection among two or more HDD's in a common enclosure? I guess that depends on how the enclosure's smarts intervene to pretend it is one drive with several partitions, or some such subterfuge.
 

Blain

Lifer
Oct 9, 1999
23,643
3
81
Originally posted by: Paperdoc
I saw some performance benchmarks recently for an external enclosure with only one HDD. The actual measured data transfer rate was almost identical for USB2 and Firewire 400 connections.
Just FYI. As I read notes, the major difference between USB2 and Firewire connections is that USB2 may be subject to more short-term interruptions in data flow.
Got any links to that "USB same as Firewire" benchmark?
All I can find is... "USB 2.0 is much slower than FireWire 400 and 800" :roll:

 

Jaxidian

Platinum Member
Oct 22, 2001
2,230
0
71
twitter.com
With Firewire 800, I don't know if my lappy supports it but I want it, as you said, for future use as well as I'm sure I will eventually have a laptop to support it.

As for USB vs Firewire you hit the nail dead-on (even assuming the two are on average comparable which I don't think is the case) - Firewire is better for steady streams of usage. I'm running virtual machines on this external drive which means I'll pretty much be accessing it non-stop. And keep in mind the CPU overhead that usually comes with a USB connection that isn't there with the Firewire connection.

Again, I'd love to go eSATA if somebody can recommend to me a good PC Card that does this. I have absolutely 0 experience in this.

Thanks all for the help so far! Old Hippie, I'm about to check out your links - thanks! :)
 

Jaxidian

Platinum Member
Oct 22, 2001
2,230
0
71
twitter.com

Paperdoc

Platinum Member
Aug 17, 2006
2,381
310
126
Sorry, it appears I mentally mashed two or more items together here. But these links will provide the answers. For reference, design maximum data transfer speeds are claimed to be:
USB2: 480 Mb/s = 60 MB/s by my calcs
Firewire 400: 400 Mb/s = 50 MB/s
Firewire 800: 800 Mb/s = 100 MB/s
eSATA / SATAII: 300 MB/s

IcyDock MB559US-1S External Enclosure Performance Review
The enclosure has USB2 and eSATA ports. Measured data transfer rate USB2 was 32 MB/s Read, 27 MB/s Write. Measured on eSATA ranged from 61 down to 31 MB/s (read and write the same), depending on how full the disk was.

Seagate ST3500601XS-RK external HD performance review
Contains benchmark data showing these trends in actual data transfer rates:
Firewire 800: Read 75 MB/s, Write 55 MB/s
Seagate eSATA: Read 49 MB/s, Write 48 MB/s
Seagate Internal: Read 48 MB/s, Write 48 MB/s (same drive, but direct internal connected)
Firewire 400: Read 38 MB/s, Write 30 MB/s
USB2: Read 31 MB/s, Write 25 MB/s

So I mis-spoke, if you will. Firewire 400 (that is, IEEE 1394a) IS a bit faster than USB2 despite the design max. In fact, both only deliver about half their max ratings - not surprising. Firewire 800 (IEEE 1394b) actually does deliver almost double the performance of the basic Firewire 400. And eSATA, despite its design max, falls in between the two versions of Firewire, but certainly well ahead of Firewire 400 (IEEE 1394a).

By the way, many people have claimed that, although SATAII drives can work up to 300 MB/s, very few current mobo's and controllers are up to this, and most will deliver performance only up to the original SATA speeds max of 150 MB/s. Until, of course, better controllers are implemented widely. Now, consider that the IcyDock review disclosed that the max actual measured data transfer rate (read and write the same) started out at 61 MB/s for an empty drive, then declined smoothly to 31 MB/s on the full drive. AVERAGE for the Seagate eSATA unit was 48 MB/s. Not too far off the Firewire 800 systems which claim 100 MB/s and deliver 75 (read) and 55 (write).

In some reviews and comments people will say, "Oh well, it may be slightly slower but nobody would actually use an external drive to run from, so it does not matter." Note that, in the Seagate external review, there was NO difference in data transfer performance for the same drive connected either via eSATA or via the "normal" internal direct connection. Impressive for eSATA, no?
 

Jaxidian

Platinum Member
Oct 22, 2001
2,230
0
71
twitter.com
Thanks for the impressive research! So it sounds like for now, Firewire 800 is the way to go, even over eSATA.

A little off-topic...
While SATA is far below it's theoretical maximum, I wonder if that is really SATA's fault. For instance, look at IDE. You have IDE 133 out there. But no way in hell are you getting 133MB/s. IDE has had plenty of time to mature and it's still far under that theoretical maximum. So does this mean that the hard drives simply cannot push that much data regardless of the interface? I suspect this is the case.
 

Jaxidian

Platinum Member
Oct 22, 2001
2,230
0
71
twitter.com

Old Hippie

Diamond Member
Oct 8, 2005
6,361
1
0
Anybody have any reasons why I shouldn't go with this enclosure?
None by me. Please, do yourself a favor and inquire about the availability of this unit. These guys aren't the easiest to work with and returns are penalized 15%.
 

Aztech

Golden Member
Jan 19, 2002
1,922
0
0
I recently started looking into the eSATA interface and I thought I saw that it would be MUCH faster than USB or 1394. I'm looking for the articles now...

Here's a quote from this article: sata-io.org
The results of eSATA are dramatic and with no protocol overhead issues as with USB or 1394. The eSATA storage bus delivers as much as 37 times more performance. This ability is perfect for using an array of drives with performance striping behind the eSATA host port.

And from the same article:
Key benefits of eSATA:Up to 6 times faster than existing external storage solutions: USB 2.0, & 1394 Robust and user friendly external connection
High performance, cost effective expansion storage
Up to 2 meter shielded cables and connectors

Two vastly different numbers (6 times versus 37 times) but regardless it sounds much faster than anything else external...
 

Paperdoc

Platinum Member
Aug 17, 2006
2,381
310
126
The marketeers are still trying to confound us with hype in place of facts. I looked closely at the source of those two quotes, and they are telling the truth. eSATA has a MAXIMUM DESIGN SPEED of 300 MB/s, which is 6 times faster than the DESIGN MAX for Firewire 400 (IEEE 1394a), and 37 times faster than the older USB 1.1. All true. But when someone sets up a real-world system and actually measures the data throughput, it always comes out lower than design max's; by how much is unpredictable. That is why I find the real measurements useful. In the past there have been all sorts of comments about why Ultra ATA 133 does not come close to 133.

I am well aware that some tests can be faulty, or may be unduly impacted by the specifics of which device and manufacturer is being tested. For example, when you test Company X's SATA II HDD mounted in Company Y's external enclosure with an eSATA interface, what does that tell us about all the other manufacturers' SATA II HDD's? Sometimes the best answers are from several tests by different organizations, but that usually takes time.

Nevertheless, the tests I did cite above verify that eSATA is faster than USB2 or IEEE 1394a links, but not quite as fast as IEEE 1394b. The exact factors may be in question, but there is a clear difference.