Features - more
Iomega is finally walking the talk when it comes to incorporating cloud features into its products. You'd expect that they would support backup to sister company Mozy. But they have added Amazon S3 and EMC's Avamar backup options, too.
You no longer get copies of EMC's Retrospect client backup software, though. Iomega figures that since their NASes now back up to and from SMB/CIFS (Windows) shares and support Mac OS Time Machine that they now have client backup covered.
The other key cloud feature in the px6 is Iomega's new Personal Cloud service. My colleague Craig Ellison covered this feature in depth in his Home Media Network Hard Drive - Cloud Edition review.
Figure 6: Personal cloud
The quick story on Personal Cloud is that it's a private VPN established between a Personal Cloud enabled storage device and client software that must installed on a device (computer) to enable access to shares sitting on the Personal Cloud device.
Personal Cloud isn't quite as easy to use as, say, Pogoplug cloud sharers and requires that you open router ports manually or have a router supporting UPnP NAT traversal. It also requires a dynamic DNS service to map an Iomega-provided subdomain, but Iomega bundles that in at no charge.
Another welcomed change is that Iomega now has an SDK that can be used to build apps for its Lifeline NASes. Unfortunately, there are no apps yet to be found at http://lifelineapps.com/.
Figure 7: Application Manager
I tested the px6-300d after upgrading its firmware to 126.96.36.199882, using our NAS test process to run tests with RAID 0, 5 and 10 volumes using all six drives.
The Benchmark summary (Figure 8) shows a significant difference between Windows and NASPT File copy tests. For example, RAID 5 Windows file copy measured only 25 MB/s, while RAID 5 NASPT file copy came in at 80 MB/s! You can see the same scenario played out in the RAID 0 and 10 tests, too.
Figure 8: px6-300d Benchmark Summary
I thought the difference could be due to write cacheing, which is set to Enabled with UPS by default (the other settings are Always Enabled and Always Disabled). But since cacheing was disabled for both Windows filecopy and NASPT tests, the cause must lie elsewhere.
As I typically find, NASPT Directory Copy tests stayed in the mid single digits.
I ran backup tests using the USB 2 and USB 3 ports with an Iomega Ultramax Pro drive configured in RAID 0. I switched back from the Buffalo drives I had been trying out because I found I needed the RAID 0 performance to fully test eSATA-equipped NAS backup.
There is no built-in formatter, so I could run only FAT32 and NTFS formatted tests. As the summary above shows, backup to a FAT formatted USB 3.0 drive was fastest at 74 MB/s.
Rsync backup to the NAS testbed running Delta Copy came in at a respectable 37 MB/s.
Please note that backup performance may be over or understated because the log timestamps used in the calculation are only to the nearest minute vs. 100ths of seconds in all other products.
iSCSI write and read to a 10 GB target on a six-drive RAID 5 volume measured a chart-topping 100 MB/s for write, but only a middling 67 MB/s for read.
I usually filter the comparative RAID 5 File Copy Write and Read charts to show products with the same number of drive bays. But with only one other six-bay product (Cisco's NSS326D06-K9), that was a boring chart.
Instead, I decided to use NASPT RAID 5 File Copy benchmarks to compare the px6-300d. This limits the comparison to newer NASes, but does show a mix of drive bay capacities. For reference when looking at the charts, the other D525-based NASes are the QNAP TS-459 Pro+, QNAP TS-559 Pro II and Thecus N5200XXX.
For NASPT RAID 5 File Copy Write (Figure 9), the px6 is clearly outclassed by all the other D525 NASes.
Figure 9: RAID 5 NASPT File Copy Write Comparison
The story is similar for NASPT RAID 5 read (Figure 10), although the spread is tighter.
Figure 10: RAID 5 NASPT File Copy Read Comparison
Iomega also sent along a couple of SSD's (128 GB Micro RealSSD C400 MTFDDAC128MAM). I didn't really expect to see a big performance boost because in NASes there are too many other factors that dominate performance.
I ran our test suite using a single drive, with the results shown in Table 1
|Windows File Copy Write||74.36|
|Windows File Copy Read||75.15|
Table 1: SSD performance test results
See any spectacular performance bump from using an SSD? Me neither.
I'm glad to see Iomega continuing to improve its products and the new Atom D525 based px4's and px6 have a lot to like. They are now second only to NETGEAR ReadyNASes in backup flexibility and the S3 cloud backup option is a welcome addition to Mozy. The ability to configure multiple storage pools and volumes within those pools is very flexible and iSCSI write performance is surprisingly high.
It's good too, to see diskless offerings and an add-in feature. Given Iomega's wide distribution and brand name, I'd think that app developers would be drawn to the platform. But the current lack of any apps and absence of information on the SDK makes this only a "future" for the moment.
As for value, unfortunately our Price vs. Performance chart is no help, since the supplied model is with six drives, which pushes up the cost. But comparing a diskless px6-300d for $1050 against a naked QNAP TS-659 Pro II at around $1340 seems to indicate that Iomega isn't trying to push the price envelope as hard.
Unfortunately, Iomega isn't pushing the performance envelope as hard as QNAP, either. This definitely forces buyers into a classic price vs. performance quandry, with only Iomega's three year warranty (with registration) vs. QNAP's 1 year to help tip the scales.