Will you be doing any testing of these devices and comparing them to other external drives so we can see if the speed gained from TB over USB3.0 of eSATA is worth any difference in price?
There is no speed gain over eSATA since these drives all are SATA internally, so there is no way that these drives can be faster than a native SATA 300 port.
And there is probably no speed gain over USB3 either since the largest number out there (327MB/s) is only about half of USB3's speed. The performance of controllers, however, could go on either side.
In the end, this is the maximum burst read speed. Average sustained speed is probably more like 150MB/s or so.
Remember that 2big has two 3.5" hard drives in RAID 0 (or RAID 1 or JBOD if you want), thus sustained 327MB/s sounds reasonable. eSATA 6Gb/s would be sufficient for that but eSATA 3Gb/s would bottleneck it a bit.
Most external enclosures, regardless of the number of drives they can hold, only offer a single eSATA port for connection to the host. So yes, it is quite common to connect 2 or more SATA drives within an external enclosure to a host using a single eSATA cable.
The test that would provide the data the original poster was hoping for could be done by pitting the LaCie 2big Thunderbolt series up against the 2big USB 3.0 and 2big Quadra using eSATA. The 2big Quadra would no doubt be the slowest of the pack seeing as it only offers an eSATA 3 Gbit/s connection. I actually just did a quick search for a 2-bay eSATA 6 Gbit/s enclosure, and surprisingly was unable to find a single one.
The fastest 7200rpm SATA hard drives do not even get 150MB/s when doing sequential readings. That's why I don't really believe the 327MB/s numbers, unless they count cache reading speed.
In the real life I don't think anyone would notice the SATA 3Gb/s bottleneck, especially since a lot of stuff is actually random access instead of sequential. And we all know Lacie will ask twice as much for this Thunderbolt solution, so it's not going to be worth it.
Under the "fastest performance", the drive achieves up to 163MB/s. Of course the numbers LaCie claims are the absolute maximums but judging by that review, they are somewhat reasonable. I think ~300MB/s in real world sounds plausible,
I'd be willing to wager LaCie is using the 1TB per platter Seagate drives, which would make their numbers even more plausible—193.0/191.6 MB/s max sequential read/write, 164.96/165.17 MB/s average. http://www.storagereview.com/seagate_barracuda_3tb...
As to whether it's worth it vs. USB 3.0 or eSATA 6 Gbit/s, not if you only look at it from a price/performance perspective. But from a practical standpoint, if you own a 2011 Mac, you have Thunderbolt but you don't have the other two interfaces, so yes, it is worth it. The 2011 17-inch MacBook Pro at least has an ExpressCard/34 slot, so you could get similar performance from a solution such as the FirmTek SeriTek/6G eSATA adapter, but that is not an inexpensive proposition either at $119.95.
While 327 MB/s would probably be achievable over USB 3.0 using the same pair of HDD's, the highest number I've ever seen posted in an AnandTech article for single controller sequential USB 3.0 throughput is 326 MB/s (http://www.anandtech.com/show/4688/highpoints-rock... The highest throughput claim that Renesas (NEC) has made for their latest USB 3.0 controller and UAS driver is only 370 MB/s. Even the USB-IF indicates that the upper limit for real world USB 3.0 throughput is right around 400 MB/s. 654 (twice 327) is a significantly larger number than 400 (the maximum speed of USB 3.0).
This particular device is limited by the HDD's used. The Thunderbolt controller is good for 1000 MB/s, and (if it is indeed using the same chip as the Thunderbolt LBD) the SATA controller is a Marvell 88SE9182, 2-port, 6 Gbit/s controller capable of 490 MB/s sequential reads on a single port (http://www.legionhardware.com/articles_pages/gigab... Unfortunately, the other limiting factor is the lack of Mac OS X driver support for the 88SE9182, which only allows 3 Gbit/s link speeds to be negotiated for the attached drives.
Thus as Kristian pointed out, the result is similar to connecting two drives in RAID 0 via a single eSATA 6 Gbit/s port.
You said, "(327MB/s) is only about half of USB3's speed." 327 is exactly half of 654, whereas 200 would be half of USB 3.0's maximum potential throughput. 327 is almost exactly 100% of USB 3.0's real world throughput achievable with currently shipping controllers and drivers.
If you meant the speed of the individual drives is about half of USB 3.0's speed, then yes, I totally agree.
Sorry, I've fixed it now. In theory, a round also means a circle so it's not wrong. Rounds per minute is more often used with guns while revolutions per minute is used with hard drives.
In my current computer: Dell T7500 LSI 9260-8i 2 x 1TB 7200RPM SAS Drives - RAID1 - OS 2 x 2TB 7200RPM SAS Drives - RAID0 - Temp Data 3 x 2TB 7200RPM SAS Drives - RAID5 - Data Storage
Throughput on the RAID0 SAS Drives using Seq tests: 397.2 Read MB/S 294.7 Write MB/s
Based on these numbers, I'm impressed by the Lacie unit... But I would rather see a RAID1 or 5 for such a large capacity of data.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
23 Comments
Back to Article
solipsism - Wednesday, February 15, 2012 - link
Will you be doing any testing of these devices and comparing them to other external drives so we can see if the speed gained from TB over USB3.0 of eSATA is worth any difference in price?zorxd - Wednesday, February 15, 2012 - link
There is no speed gain over eSATA since these drives all are SATA internally, so there is no way that these drives can be faster than a native SATA 300 port.And there is probably no speed gain over USB3 either since the largest number out there (327MB/s) is only about half of USB3's speed. The performance of controllers, however, could go on either side.
In the end, this is the maximum burst read speed. Average sustained speed is probably more like 150MB/s or so.
MySchizoBuddy - Wednesday, February 15, 2012 - link
where did eSATA come in?Kristian Vättö - Wednesday, February 15, 2012 - link
Remember that 2big has two 3.5" hard drives in RAID 0 (or RAID 1 or JBOD if you want), thus sustained 327MB/s sounds reasonable. eSATA 6Gb/s would be sufficient for that but eSATA 3Gb/s would bottleneck it a bit.zorxd - Wednesday, February 15, 2012 - link
Except that you don't plug 2 SATA hard drives on a single port.You use two ports, so SATA 300 is more than enough.
repoman27 - Wednesday, February 15, 2012 - link
Most external enclosures, regardless of the number of drives they can hold, only offer a single eSATA port for connection to the host. So yes, it is quite common to connect 2 or more SATA drives within an external enclosure to a host using a single eSATA cable.The test that would provide the data the original poster was hoping for could be done by pitting the LaCie 2big Thunderbolt series up against the 2big USB 3.0 and 2big Quadra using eSATA. The 2big Quadra would no doubt be the slowest of the pack seeing as it only offers an eSATA 3 Gbit/s connection. I actually just did a quick search for a 2-bay eSATA 6 Gbit/s enclosure, and surprisingly was unable to find a single one.
Zak - Thursday, February 16, 2012 - link
What if I want to plug 4 or more drives in RAID and I don't want four or more separate enclosures and max possible speed?repoman27 - Thursday, February 16, 2012 - link
In that case, get a Pegasus: http://www.promise.com/storage/raid_series.aspx?m=...zorxd - Thursday, February 16, 2012 - link
If you want 4 drives I say get a desktop.zorxd - Thursday, February 16, 2012 - link
According to http://www.anandtech.com/show/5042/seagates-new-ba...The fastest 7200rpm SATA hard drives do not even get 150MB/s when doing sequential readings. That's why I don't really believe the 327MB/s numbers, unless they count cache reading speed.
In the real life I don't think anyone would notice the SATA 3Gb/s bottleneck, especially since a lot of stuff is actually random access instead of sequential.
And we all know Lacie will ask twice as much for this Thunderbolt solution, so it's not going to be worth it.
Kristian Vättö - Thursday, February 16, 2012 - link
See this: http://www.storagereview.com/hitachi_deskstar_7k30...Under the "fastest performance", the drive achieves up to 163MB/s. Of course the numbers LaCie claims are the absolute maximums but judging by that review, they are somewhat reasonable. I think ~300MB/s in real world sounds plausible,
repoman27 - Thursday, February 16, 2012 - link
I'd be willing to wager LaCie is using the 1TB per platter Seagate drives, which would make their numbers even more plausible—193.0/191.6 MB/s max sequential read/write, 164.96/165.17 MB/s average. http://www.storagereview.com/seagate_barracuda_3tb...As to whether it's worth it vs. USB 3.0 or eSATA 6 Gbit/s, not if you only look at it from a price/performance perspective. But from a practical standpoint, if you own a 2011 Mac, you have Thunderbolt but you don't have the other two interfaces, so yes, it is worth it. The 2011 17-inch MacBook Pro at least has an ExpressCard/34 slot, so you could get similar performance from a solution such as the FirmTek SeriTek/6G eSATA adapter, but that is not an inexpensive proposition either at $119.95.
repoman27 - Wednesday, February 15, 2012 - link
While 327 MB/s would probably be achievable over USB 3.0 using the same pair of HDD's, the highest number I've ever seen posted in an AnandTech article for single controller sequential USB 3.0 throughput is 326 MB/s (http://www.anandtech.com/show/4688/highpoints-rock... The highest throughput claim that Renesas (NEC) has made for their latest USB 3.0 controller and UAS driver is only 370 MB/s. Even the USB-IF indicates that the upper limit for real world USB 3.0 throughput is right around 400 MB/s. 654 (twice 327) is a significantly larger number than 400 (the maximum speed of USB 3.0).This particular device is limited by the HDD's used. The Thunderbolt controller is good for 1000 MB/s, and (if it is indeed using the same chip as the Thunderbolt LBD) the SATA controller is a Marvell 88SE9182, 2-port, 6 Gbit/s controller capable of 490 MB/s sequential reads on a single port (http://www.legionhardware.com/articles_pages/gigab... Unfortunately, the other limiting factor is the lack of Mac OS X driver support for the 88SE9182, which only allows 3 Gbit/s link speeds to be negotiated for the attached drives.
Thus as Kristian pointed out, the result is similar to connecting two drives in RAID 0 via a single eSATA 6 Gbit/s port.
zorxd - Wednesday, February 15, 2012 - link
327 is the combined maximum speed of both drives in RAID0. A single drive should achieve about half of that. There is no 654 in the equation.repoman27 - Wednesday, February 15, 2012 - link
You said, "(327MB/s) is only about half of USB3's speed." 327 is exactly half of 654, whereas 200 would be half of USB 3.0's maximum potential throughput. 327 is almost exactly 100% of USB 3.0's real world throughput achievable with currently shipping controllers and drivers.If you meant the speed of the individual drives is about half of USB 3.0's speed, then yes, I totally agree.
tech6 - Wednesday, February 15, 2012 - link
7200 "rounds per minute" - are they selling 7200 revolutions per minute drive or a really fast machine gun?Kristian Vättö - Wednesday, February 15, 2012 - link
Sorry, I've fixed it now. In theory, a round also means a circle so it's not wrong. Rounds per minute is more often used with guns while revolutions per minute is used with hard drives.MrMJ - Wednesday, February 15, 2012 - link
In my current computer:Dell T7500
LSI 9260-8i
2 x 1TB 7200RPM SAS Drives - RAID1 - OS
2 x 2TB 7200RPM SAS Drives - RAID0 - Temp Data
3 x 2TB 7200RPM SAS Drives - RAID5 - Data Storage
Throughput on the RAID0 SAS Drives using Seq tests:
397.2 Read MB/S
294.7 Write MB/s
Based on these numbers, I'm impressed by the Lacie unit... But I would rather see a RAID1 or 5 for such a large capacity of data.
Kristian Vättö - Wednesday, February 15, 2012 - link
RAID 1 is supported. RAID 5 would require more drives and a RAID controller, hence it's not supported (though you can look into Pegasus for that).MrMJ - Wednesday, February 15, 2012 - link
I wonder if JBOD would be possible with these.Kristian Vättö - Wednesday, February 15, 2012 - link
That's supported as well (take a look at the source link at the end of the article). You can do all these in Disk Utility anyway.zorxd - Wednesday, February 15, 2012 - link
Your numbers are real numbers. Lacie numbers are theoretical. In practice expect much less.hingfingg - Thursday, February 16, 2012 - link
** {{w w w }} {{proxy4biz }} {{ com}} *****