Seagate BlackArmor NAS 440 NAS Device

Article Index

Performance



In order to test the functionality of the 440, we placed it on a Gigabit-Ethernet network and accessed it from a variety of Windows and Mac desktops and laptops. We connected to it from clients via both wired and wireless connections, as well as remotely over the Internet. To test the device's performance, we used a combination of synthetic benchmark testing and real-world file copy tests. Throughout the tests, the 440 was configured with all four of its drives as a single RAID 5 volume.



Our first test was conducted with the synthetic ATTO Disk Benchmark. We mapped the 440 as drive letter Z: on the test machine and ran the default ATTO test. On the test, the 440 performed best with block sizes that are 32K when reading and writing. The fastest read speed the 440 put in on this test was 55.9MB/Sec, and the fastest write speed was 31.5MB/Sec. Seagate claims that the 440 is capable of a read transfer rate of up to 50MB/Sec when the device is configured in RAID 5 mode. It seems that at least as far as this synthetic test goes, the performance we measured actually exceeds Seagate's claim. The question is, will that claim still hold up with real-world file transfer performance? Let's see...

Our next set of test are made up of real-world data-transfer tests to and from the device over our network using an HP Pavilion Elite m9550f desktop (2.5GHz Intel Core 2 Quad Q9300, 8GB PC2-6400 DDR2 SDRAM, 1TB NTFS 7200RPM SATA hard drive, ATI Radeon HD 4850 512MB, Windows Vista Home Premium SP1 64-bit) via a Gigabit-Ethernet connection. We conducted large-file write and read tests where we copied a 1.7GB ISO file between the 440 and the m9550f. We also conducted small-files write and read tests where we copied a 267MB folder made up of 70 JPGs ranging in size from 2.27MB to 4.38MB between the 440 and the m9550f. We conducted these tests by dragging-and-dropping the folders and files in Windows, with the 440 connected as a mapped drive.

We compared the performance of the 440 against that of a number of NAS devices we've looked at, including the WD My Work World Edition, Maxtor Central Axis Business Edition, Linksys by Cisco Media Hub, HP MediaSmart Server LX195, Addonics NAS Adapter, Pogoplug, and the Synology Disk Station DS409+. We also repeated all of our tests on an older 500GB Maxtor Shared Storage NAS device--copying files between the Maxtor Shared Storage device and the m9550f. Additionally, we ran our tests on an external hard drive connected directly to a USB 2.0 port on the m9550f; the drive we used was a 320GB Western Digital Caviar Blue drive (7200RPM SATA-II, 16MB cache) placed into an external enclosure and formatted using the NTFS file system.



When transferring large files, the 440 is one of the faster NAS devices we've seen. In fact, only the Synology DS409+ and the HP MediaSmart Server LX195 show speedier write performance, and the 440 is actually a bit faster than the LX195 when reading large files (based on the tests run on the NAS devices tested on this particular test setup). The 440's write speed of 21.1MB/Sec equates to a transfer rate of 177.1Mb/Sec, and the read speed of 46.6MB/Sec is equivalent to 391.1Mb/Sec. As these transfer rates indicate, you need a Gigabit network to fully appreciate the performance potential of the 440--a 100Mb/Sec network will present a bottleneck that will limit what the 440 is fully capable of. Our measured read speed of 46.6MB/Sec is only a bit slower than Seagate's 50MB/Sec claim.



As is typically the case for NAS devices, the 440's small-file transfer performance is not as fast as it is when transferring large files. Even so, the 440's small-files write speed of 15.6MB/Sec (130.6Mb/Sec) and read speed of 28.4MB/Sec (237Mb/Sec) are still some of the speediest throughput rates we've seen. Only the Synology DS409+ is faster--albeit, noticeably faster.  The LX195 was only a hair slower than the 440 with writing and reading small files.







As the 440 also allows you to aggregate the two LAN ports in order to supposedly get faster performance, we ran some additional tests with both LAN ports active and connected to our Gigabit router. The 440 supports two different aggregation modes: Round Robin and Fail Over LAN. We expected to see faster performance in Round Robin mode than we saw when using a single LAN connection, but this was not the case. In fact, performance in Round Robin mode was actually marginally slower than what we saw with just a single LAN port in use. We queried Seagate about this surprising result, and this was the response we got:

"The best performance on the NAS 440 can be configured without Link Aggregation enabled. It is a case where the processor power in the NAS 440 does not see any real benefit for most users when using Round Robin (aggregated performance).  That's likely why you're not seeing the performance measures you may have expected in that mode."

In other words, the device is not powerful enough to take advantage of the feature--which begs the question, why offer this configuration setting at all as an option? As to the Fail Over LAN mode being slower than single LAN connection, this makes sense as the device requires additional overhead to monitor the LAN ports and switch over when necessary, while not necessarily utilizing both ports simultaneously.

Accessing the 440's iTunes server is fairly speedy: it took just over 8 seconds to access a 57.69GB library of over 6,400 songs. This was not quite as fast, however, as the mere 2 seconds it took the DS409+. Also note that you cannot password protect access to the 440's iTunes library, so any iTunes client that can see the 440 on the local network will be able to access the library.

To get a sense of how much power the 440 consumes, we connected it to a power meter. When the 440 is sitting idle, it uses around 42-watts of power. When under load, its power consumption appears to top out at around 46-watts. When in HDD Standby mode, the device uses about 14 watts of power.


Related content

Comments

Show comments blog comments powered by Disqus