We configured three of the WD Red Pro drives in a RAID-5 volume in the QNAP TS-EC1279U-SAS-RP. A CIFS share in the volume was subject to some IOMeter tests with access from up to 25 VMs simultaneously. The following four graphs show the total available bandwidth and the average response time while being subject to different types of workloads through IOMeter. IOMeter also reports various other metrics of interest such as maximum response time, read and write IOPS, separate read and write bandwidth figures etc. Some of the interesting aspects from our IOMeter benchmarking run are available here.
We see that the sequential accesses are still limited by the network link, but, this time, on the NAS side. The 100% sequential reads show similar results for all the drives. However, the WD Red Pro shows the best stability under stress from multiple clients for the 50% sequential reads / 50% sequential writes test. On the other hand, the Random 8K 70% reads sequence for the WD Red Pro show better numbers compared to the WD Red, but can't compete with the numbers from the other 6 TB drives. The 60% Random / 65% Reads sequence shows the WD Red Pro in better light compared to the WD Red and the HGST Deskstar NAS, but the other drives show consistently better numbers.
Except for upper end enterprise it's still probably too early to add 10gbe to a nas test suite.
A few years back I remember reading that there was a process threshold that a lot of people were expecting to dramatically drive down the high power cost of sending a 10gb signal over twisted pair connections. Unfortunately I don't remember if it was 14nm; in which case we should start to see lower power hardware within a year or 10 nm; in which case we'll probably end up waiting until closer to 2020 before getting in in our boxes.
Yes, that's absolutely true, my nas waste same power on the 10gbe adapter and on cpu and that's crazy. Still, with HD and then 4k videos, raw pics near or over 50MB each, old good hdd disks which continue to improve, 1gbe is going to be too limiting... 2020 is really far away! ☺
We could well see copper Ethernet being replaced by then. IBM just announced in May a 100 Gb/s transceiver using the first fully integrated wavelength multiplexed chip. The optical components are on-chip along side the electronic components. This integration is what will bring the price of optical Ethernet down to parity with copper. It will be a huge bump in bandwidth and power reduction. It is telling that the 100G consortium is already increasing the max distance to 1 km.
I've been using the Western Digital's NAS with these drives, WD EX2100, since April. Very solid unit. Really easy to set up, and includes advanced features when needed such as link aggregation.
No, they are right - 10 in 10^15 is exactly the same as 1 in 10^14. It's written/marketed that way to confuse and look much better than it really is.
Using a less than sign doesn't really change the base order of magnitude - eg. 9 in 10^15 would be consistent with their PDF table claim and it is still almost 10 times worse than any enterprise drive at 1 in 10^15 URE, which is why non-enterprise are not worth buying at these fail-likely sizes if you value your data.
but if you getting unrecoverable read errors at that point you should replace the HDD as there is a problem with the drive, like i need to replace my segate {as the theme goes with seagate drives} as its slowly developing reallocated sectors, surprisingly its still working fine
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
62 Comments
View All Comments
DanNeely - Tuesday, September 8, 2015 - link
Except for upper end enterprise it's still probably too early to add 10gbe to a nas test suite.A few years back I remember reading that there was a process threshold that a lot of people were expecting to dramatically drive down the high power cost of sending a 10gb signal over twisted pair connections. Unfortunately I don't remember if it was 14nm; in which case we should start to see lower power hardware within a year or 10 nm; in which case we'll probably end up waiting until closer to 2020 before getting in in our boxes.
Hannibal80 - Wednesday, September 9, 2015 - link
Yes, that's absolutely true, my nas waste same power on the 10gbe adapter and on cpu and that's crazy. Still, with HD and then 4k videos, raw pics near or over 50MB each, old good hdd disks which continue to improve, 1gbe is going to be too limiting...2020 is really far away! ☺
Jaybus - Wednesday, September 9, 2015 - link
We could well see copper Ethernet being replaced by then. IBM just announced in May a 100 Gb/s transceiver using the first fully integrated wavelength multiplexed chip. The optical components are on-chip along side the electronic components. This integration is what will bring the price of optical Ethernet down to parity with copper. It will be a huge bump in bandwidth and power reduction. It is telling that the 100G consortium is already increasing the max distance to 1 km.Sivar - Tuesday, September 8, 2015 - link
I've been using the Western Digital's NAS with these drives, WD EX2100, since April. Very solid unit. Really easy to set up, and includes advanced features when needed such as link aggregation.lundgsi2@yahoo.co.uk - Tuesday, September 8, 2015 - link
thats great can't wait to get my hands on thesewww.hardwarecomputerstore.co.uk
ex_User - Wednesday, September 9, 2015 - link
I believe you have an error in specs. "Non-Recoverable Read Errors / Bits Read" for Red Pro is "<10 in 10^15", not 10^14.http://www.wdc.com/wdproducts/library/SpecSheet/EN...
asmian - Thursday, September 10, 2015 - link
No, they are right - 10 in 10^15 is exactly the same as 1 in 10^14. It's written/marketed that way to confuse and look much better than it really is.Using a less than sign doesn't really change the base order of magnitude - eg. 9 in 10^15 would be consistent with their PDF table claim and it is still almost 10 times worse than any enterprise drive at 1 in 10^15 URE, which is why non-enterprise are not worth buying at these fail-likely sizes if you value your data.
ex_User - Thursday, September 10, 2015 - link
Well, that misleading marketing works very well on laymen like me -- only after your post have I noticed the <1/<10 trick.leexgx - Saturday, September 12, 2015 - link
i agree that <1/<10 is Very misleading, so they have reduced or not added more error correction on these drivesleexgx - Saturday, September 12, 2015 - link
but if you getting unrecoverable read errors at that point you should replace the HDD as there is a problem with the drive, like i need to replace my segate {as the theme goes with seagate drives} as its slowly developing reallocated sectors, surprisingly its still working fine