Netgear ReadyDATA 5200 review
Delivers top performance and extensive data protection features, but its low deduplication ratio limits storage savings
Review Date: 10 Jan 2013
Reviewed By: Dave Mitchell
Price when reviewed: £5,499 (£6,599 inc VAT)
Features & Design
Value for Money
Netgear’s ReadyNAS appliances are popular network storage choices for SMBs, but their limited features and multimedia baggage makes them less appealing to larger businesses. With the ReadyDATA 5200, Netgear aims to remedy these shortcomings, giving the old RAIDiator software the push in favour of a new Solaris-based operating system.
With the ZFS file system on board, the ReadyDATA provides NAS and IP SAN operations, unlimited snapshots and thin provisioning. There’s also built-in deduplication, compression, fast copy-on-write snapshots and SSD caching.
Block-level replication for volumes and snapshots is also included. This uses the hosted service and management portal as used in Netgear's ReadyNAS systems, and snapshots can be taken regularly, with only the available storage limiting the number and size of snapshots you can take. For each share and LUN, you can select the continuous protection option box and go for hourly, daily or weekly snapshots. The one major restriction is that it’s only possible to replicate with other ReadyDATA appliances.
The drive backplane supports 6Gbits/sec drive speeds and there are two Gigabit ports plus a dual-port 10GbE PCI Express card. Storage expansion is now possible, with two external SAS ports supporting Netgear’s EDA2000 and EDA4000 expansion chassis.
The main web interface is nicely designed and laid out. For volume creation, choose drives from the front-panel graphic and pick a RAID type; deduplication and compression can be enabled or disabled for each share and iSCSI LUN; and volume expansion allows new drives in the extra shelves to be added to existing RAID arrays. One smart feature is the ability to add extra drives to an array as read or write caches to improve performance.
The review system was supplied with six SATA and four SAS hard disks, plus Intel and STEC SSDs. We created separate RAID5 arrays on the SAS and SATA drives, and found IP SANs easy to create: choose a target name, a size, and thick or thin provisioning, and the ReadyDATA does the rest. LUNs can be expanded into available volume space and migrated online to another volume.
For performance testing, we used the lab’s Dell PowerEdge R820 with its quartet of 2.2GHz E5-4607 Xeons and 96GB of DDR3 memory. This was running Windows Server 2008 R2 Enterprise and was equipped with an Emulex 10GbE adapter.
IP SAN testing over a single 10GbE link saw Iometer report sequential read and write speeds of 656MB/sec and 540MB/sec for a 200GB target on the SAS array. Running the same test on the SATA array saw lower speeds of 620MB/sec and 510MB/sec.
We then added the STEC SSD to the SATA array as a write cache and the Intel SSD as a read cache to see what difference it would make. We saw a significant speed boost, with Iometer reporting similar speeds to the SAS array. Using an Iometer database load comprising 16KB random reads and writes saw 480MB/sec on the SAS array. The SATA array returned 455MB/sec, and enabling SSD caching again saw this boosted in line with SAS speeds.
Deduplication is carried out at the appliance so it’s transparent to all host systems. To test it we used our deduplication test suite and CA’s ARCserve r16, set to use a mapped share as a disk-based backup. Using 4GB of 1,000 files, we ran a standard backup strategy consisting of daily incrementals and full weekly backups.
After the first full backup had run, 2% of data was modified in 40% of files prior to each subsequent backup. After a four-week simulation, we saw a reduction ratio of 3.04:1. This is very low, but on par for ZFS-based systems. Infortrend’s EonNAS Pro 500, another ZFS appliance, returned 3.2:1 in the same test.
The ReadyDATA 5200 is a better bet for business storage than the ReadyNAS boxes since it has superior features and data protection options. Performance is better as well, although the low data reduction ratios mean it won’t save as much storage as most other appliances boasting deduplication.
Author: Dave Mitchell
We have a ReadyDATA with 6 Seagate SATA drives for a while. We created a RAID 5 iSCSI volume and ran IOmeter sequential tests. It came out that seq read is only 133MB/s and seq write is 245MB/s. Could you provide details of how to test it?
By Swlin on 18 May 2013
We used 4 disk workers assigned to the iSCSI target, 10 o/s I/Os, 256KB transfer request sizes and 100 per cent sequential read or write distributions.
By DaveMitchell on 18 May 2013
Random IO - Iometer
Thanks for the hints. We also wonder the Iometer setting for 16KB random test. The result of 455MB/s for 16KB random IO is absolutely great. Is it 100% random read or 100% random write or a different weight ratio? From this result, the IOPS of 16KB random IO can fetch to 28K. It's pretty good.
By Swlin on 20 May 2013
Random IO - Iometer
For SQL database performance testing, we used 256 outstanding I/Os, a 16KB block size, 66 percent read, 34 percent write and a 100 percent random distribution.
By DaveMitchell on 20 May 2013
Thanks for your answer. May we know what setting is for "maximum disk size" in IOmeter for all these testing data? Or the size of iobw.tst file. All these test results should be non cache hit performance, am I correct?
By Swlin on 21 May 2013
- IDC: iPad intertia opens door for Windows tablets
- Chip breakthrough to eliminate checkout queues
- Rivals put on notice as Spotify snaps up The Echo Nest
- Windows 8.1 Update 1 leaks via Microsoft's website
- Bitcoin "founder" says: you've got the wrong man
- Has bitcoin creator been found?
- HTC Desire 310: more competition for the Moto G
- Mozilla questions why Dell charges £16 to install Firefox
- Getty makes millions of photos free to embed
- Roku beats Chromecast to the UK with £50 streaming stick
- Getty joins the race to the bottom
- Hour of Code: five steps to learn how to code
- Sony Xperia Z2 Tablet review: first look
- Sony Xperia Z2 review: first look
- Samsung Galaxy Gear 2 review: first look
- Nokia XL review: first look
- Samsung Galaxy S5 review: first look
- Nokia X review: first look
- Censorship by copyright: Myles Powers and abuse of DMCA takedowns
- Turn an old smartphone into an in-car entertainment system
- Old-school internet scams: five that just won't die
- Bitcoin believers not worried by Mt. Gox disarray
- How to hack your car
- Small server vs cloud: which is best for SMBs?
- Block party: why do millions play Minecraft?
- What to do if you’re still on Windows XP
- Microsoft Word: top 20 secret features
- Measuring me: is your body the future of security?
- The best mobile apps for business
- Adobe Photoshop: top 20 secret features
- Windows Server 2012 R2: how the Datacenter edition could change SMBs
- Invoices and VAT: how to set up your documents correctly
- Nexus 5 vs Samsung Galaxy S4 Active: the best phone for avoiding screen burn
- How much is a social user worth?
- The key to choosing a secure password
- Thunderbolt Bridge: a fast Mac migration tool
- Should you advertise on Twitter?
- How to track a lost smartphone
- Self-publishing success: the best way to sell your book
- 1.6TB SSD: why would you need one?