Introduction and Evaluation Methodology

NAS units with four bays present the best balance between cost and expandability for home consumers. However, with increasing hard drive sizes, two bays make the cut for many usage scenarios. The demand for high-performance, but cost-effective NAS units has been picking up, and this is where the modern ARM-based platforms come into play. Netgear's ReadyNAS 200 series was launched at the 2015 CES and sported an ARM Cortex A15-based Annapurna Labs SoC. Negear's offerings differentiates itself from the competition due to use of btrfs as the file system for the data volume. We have already looked at the Intel Atom-based RN312. In this review, we will take a look at how the ARM-based RN202 performs with the ReadyNAS OS, and see how the unit stacks up against the competitors in this space.

The specifications of the Netgear RN202 are provided in the table below

Netgear RN202 Specifications
Processor	Annapurna Labs SoC (2C/2T ARM Cortex A15 @ 1.4 GHz)
RAM	2GB
Drive Bays	2x 3.5"/2.5" SATA II / III HDD / SSD (Hot-Swappable)
Network Links	2x 1 GbE
External I/O Peripherals	3x USB 3.0, 1x eSATA
Expansion Slots	N/A
VGA / Display Out	N/A
Full Specifications Link	Netgear RN202 Specifications
Price	USD 282

The various specifications of the NAS are backed up by the data gleaned via SSH access to the unit.

The ReadyNAS 200 series also includes a 4-bay variant. The comparison between the two units (including the hardware specifications) are reproduced from Netgear's marketing material below.

The industrial design of the unit is the same as that of the ReadyNAS RN312. In terms of user experience with the hardware, the toolless drive caddies are one of the best designs we have seen across samples from most of the players in the industry. That said, they are a bit non-intuitive for first-time users, but they get the job done while also providing a bit of vibration dampening for the 3.5" drives. 2.5" drives still need screws for securing in the same caddy.

The setup process is quite straightforward. Upon connection to the network, the RN202 receives a DHCP address even in a diskless state. The IP address can be determined either from the DHCP provider in the system or via Netgear's RAIDar utility. Accessing the IP address with the default 'admin'/'password' login enables the setup process shown in the gallery below. We started off with one disk in the unit, and it was configured as a JBOD volume with X-RAID by  default. Support exists for manually defragging, scrubbing and balancing the btrfs volume. Hot-swapping of drives is possible and adding a new drive or replacing a failed drive with X-RAID enabled automatically triggers expansion / rebuild.

Similar to almost all other NAS units in the market, there is support for viewing the S.M.A.R.T attributes of the member disks. The settings section allows choice of various services to enable (SMB, AFP, NFS, FTP, SSH, ReadyDLNA, uPnP etc.). The user experience is a bit inconsistent here in terms of the interface. While clicking on the service buttons toggles the inner rectangle between green (enabled) and gray (disabled) in this section, there are other sections where similar buttons can't be clicked to toggle status. The settings page allows configuration of other aspects such as update management, backing up settings and alerts.

The logs section records the various NAS activities with timestamps and the power section enables power scheduling, disk spin-down configuration, Wake-on-LAN settings and UPS configuration.

Creating new shares allows us to configure bit-rot protection (disabled by default, results in a performance hit on ARM-based systems), compression (since btrfs provides native compression capabilities), snapshot scheduling and protocols with which it can be accessed. The web UI also features a built-in file browser for the NAS contents and includes a timeline view (based on the snapshots).

The network settings allow the interfaces to be bonded. It is possible to set up 802.3ad LACP (amongst other bonding modes). There are a number of third-party apps available (though the selection is nowhere close to what Synology and QNAP have). Some cloud management features (remote access via VPN, replication over the Internet etc.) are also available.

In the rest of the review, we will take a look at the single client performance for SMB and iSCSI, followed by a look at what enabling encryption entails. We will have three sections dealing with multi-client scenarios across a number of different client platforms as well as access protocols. Prior to all that, we will take a look at our testbed setup and testing methodology.

Testbed Setup and Testing Methodology

The Netgear RN202 can take up to 2 drives. Users can opt for either JBOD, RAID 0 or RAID 1 configurations. We expect typical usage to be with a single RAID-1 volume. To keep things consistent across different NAS units, we benchmarked a single RAID-1 volume. Two Western Digital WD4000FYYZ RE drives were used as the test disks. Our testbed configuration is outlined below.

AnandTech NAS Testbed Configuration
Motherboard	Asus Z9PE-D8 WS Dual LGA2011 SSI-EEB
CPU	2 x Intel Xeon E5-2630L
Coolers	2 x Dynatron R17
Memory	G.Skill RipjawsZ F3-12800CL10Q2-64GBZL (8x8GB) CAS 10-10-10-30
OS Drive	OCZ Technology Vertex 4 128GB
Secondary Drive	OCZ Technology Vertex 4 128GB
Tertiary Drive	OCZ Z-Drive R4 CM88 (1.6TB PCIe SSD)
Other Drives	12 x OCZ Technology Vertex 4 64GB (Offline in the Host OS)
Network Cards	6 x Intel ESA I-340 Quad-GbE Port Network Adapter
Chassis	SilverStoneTek Raven RV03
PSU	SilverStoneTek Strider Plus Gold Evolution 850W
OS	Windows Server 2008 R2
Network Switch	Netgear ProSafe GSM7352S-200

The above testbed can run up to 25 Windows 7 or CentOS VMs simultaneously, each with a dedicated 1 Gbps network interface. This simulates a real-life workload of up to 25 clients for the NAS being evaluated. All the VMs connect to the network switch to which the NAS is also connected (with link aggregation, as applicable). The VMs generate the NAS traffic for performance evaluation. However, keeping in mind the nature of this unit, we restricted ourselves to a maximum of 10 simultaneous clients.

Thank You!

We thank the following companies for helping us out with our NAS testbed:

• Thanks to Intel for the Xeon E5-2630L CPUs and the ESA I-340 quad port network adapters
• Thanks to Asus for the Z9PE-D8 WS dual LGA 2011 workstation motherboard
• Thanks to Dynatron for the R17 coolers
• Thanks to G.Skill for the RipjawsZ 64GB DDR3 DRAM kit
• Thanks to OCZ Technology for the two 128GB Vertex 4 SSDs, twelve 64GB Vertex 4 SSDs and the OCZ Z-Drive R4 CM88
• Thanks to SilverStone for the Raven RV03 chassis and the 850W Strider Gold Evolution PSU
• Thanks to Netgear for the ProSafe GSM7352S-200 L3 48-port Gigabit Switch with 10 GbE capabilities.
• Thanks to Western Digital for the two WD RE hard drives (WD4000FYYZ) to use in the NAS under test.

Single Client Performance - CIFS & iSCSI on Windows

The single client CIFS and iSCSI performance of the Netgear RN202 was evaluated on the Windows platforms using Intel NASPT and our standard robocopy benchmark. This was run from one of the virtual machines in our NAS testbed. All data for the robocopy benchmark on the client side was put in a RAM disk (created using OSFMount) to ensure that the client's storage system shortcomings wouldn't affect the benchmark results. It must be noted that all the shares / iSCSI LUNs are created in a RAID-1 volume.

The main surprise in the results below is that the RN202 performs better than the RN312 for certain access traces. This could be due to performance upgrades in the firmware. In general, the reads are very fast - easily the best amongst the NAS units we have evaluated, while writes tend to come towards the middle of the pack. The fast reads skew the numbers heavily in the read-centric workloads.

We created a 250 GB iSCSI LUN / target and mapped it on to a Windows VM in our testbed. The same NASPT benchmarks were run and the results are presented below. The observations we had in the CIFS subsection above hold true here too.

The iSCSI performance is similar to the CIFS performance. The writes suffer heavily, but the read performance is again very good enabling the unit to emerge as the clear leader in certain workloads. Note that the test share was configured with bit-rot protection (something users expect from any btrfs-enabled NAS) and a weekly snapshot schedule, but no compression was in play during the benchmarking process.

Encryption Support Evaluation

Consumers looking for encryption capabilities can opt to encrypt a iSCSI share with TrueCrypt or some in-built encryption mechanism in the client OS. However, if requirements dictate that the data must be shared across multiple users / computers, relying on encryption in the NAS is the best way to move forward. Most NAS vendors use the industry-standard 256-bit AES encryption algorithm. One approach is to encrypt only a particular shared folder while the other approach is to encrypt the full volume. Netgear supports only volume-level encryption. In addition, a USB drive (into which the key is written at the time of volume creation) needs to be attached to the unit for the encrypted volume to remain / be mounted.

On the hardware side, encryption support can be in the form of specialized hardware blocks in the SoC (common in ARM / PowerPC based NAS units). In x86-based systems, accelerated encryption support is dependent on whether the AES-NI instruction is available on the host CPU. The Annapurna Labs SoC has hardware crypto engines that enable minimal hit in performance with encrypted volumes.

The performance of the encrypted volume is easily the best amongst all the ARM-based NAS units that have been evaluated so far. The x86-based units in the list above don't have AES-NI, and hence, the ARM-based RN202 easily manages to win out. There is definitely a performance hit compared to unencrypted volumes, and this can only be resolved by going to higher performance platforms./p>

Multi-Client CIFS Performance for Consumer Workloads

The workloads experienced by a NAS unit in a typical home consumer setting have changed quite a bit over the last few years. Multiple mobile devices in a typical household raise the possibility that a NAS could be subject to the streaming out of multiple video files simultaneously. The popularity of IP cameras also make it necessary for NAS units to be able to record multiple video streams at the same time.

In our previous NAS reviews, we evaluated multi-client scenarios using synthetic workload traces and IOMeter. While there is nothing wrong in presenting numbers from such benchmarks, the reader is often left confused as to what those numbers might mean for his particular use-cases. Intel's NASPT benchmarking program gives us a good idea of the performance of the NAS unit when accessed by a single client. We took the source code of Intel's NASPT along with the supplied application traces and tweaked them to be able to run from more than one Windows client simultaneously in a co-ordinated manner. The graphs below present the results from tracking various metrics during the course of the benchmark runs. It must be noted that the average service times refer to what is obtained for all the traces when some of the data has already been cached in the client's memory. Unfortunately, NASPT doesn't provide any sort of guideline on what the optimal bandwidth and service times are for a good user experience.

Content Creation

The Content Creation workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

Folder Copy from NAS

The Folder Copy from NAS workload seems to get acceptable performance for up to 5 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

Folder Copy to NAS

The Folder Copy to NAS workload seems to get acceptable performance for up to 4 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

File Copy from NAS

The File Copy from NAS workload seems to get acceptable performance for up to 4 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

File Copy to NAS

The File Copy to NAS workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

HD Video (1x) Playback

The HD Video (1x) Playback workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

HD Video(1x) Playback and Record

The HD Video(1x) Playback and Record workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

HD Video (1x) Record

The HD Video (1x) Record workload seems to get acceptable performance for up to 4 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

HD Video (2x) Playback

The HD Video (2x) Playback workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

HD Video (4x) Playback

The HD Video (4x) Playback workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

Office Productivity

The Office Productivity workload seems to get acceptable performance for up to 5 clients. Beyond that, we have a noticeable drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

Photo Album

The Photo Album workload seems to get acceptable performance for up to 3 clients. Beyond that, we have a drop in the per-client bandwidth numbers. The detailed table with a breakdown of all the throughput numbers as well as the service times is available here

The detailed logs from the processing of our benchmarks - inclusive of metrics such as the file open times for each workload on each of the clients - can be found here

Multi-Client CIFS Performance for Professional Workloads

NAS units used in SMBs / SMEs need to provide good performance under heavy load from multiple clients. The SPEC SFS 2014 benchmark uses real-life workloads (just like Intel NASPT), but makes it easier for users to understand the benchmark results. This is achieved by using the concept of business metrics. Given a particular NAS unit, how many concurrently accessed databases can reside in it? How many IP cameras or video streams can be simultaneously recorded? To determine this metric, each load point is associated with a target required op rate. If the NAS under test doesn't meet that op rate, it is deemed as an 'invalid run'. SPEC requires all published benchmarks to follow certain strict rules - such as presented results having no invalid runs for at least 10 load points. Unfortunately, small-scale NAS systems with 7200 RPM drives can't meet these requirements, Hence, we can't officially publish SPEC SFS 2014 benchmark results for the evaluation of the Netgear ReadyNAS RN202.

Using a popular filer benchmarking program, we did play back multi-client real-world professional workload access traces on the NAS using up to 10 Windows 7 VMs. However, failing to meet the required op rate criteria at a particular load point made us stop the testing a couple of load points down the road. The Netgear ReadyNAS RN202 with two 7200 RPM hard drives in RAID-1 can support recording of more than 10 video streams, up to 3 databases and at the most, 2 virutal machines.

Database Operations

The Database Operations workload seems to get acceptable performance for up to 3 clients.The detailed metrics from our trace playback are available here

Software Builds

The Software Builds workload doesn't seem to get acceptable performance for even one client. The detailed metrics from our trace playback are available here

Video Recording

The Video Recording workload gets acceptable performance for up to 10 clients (the maximum we tested).The detailed metrics from our trace playback are available here

Virtual Desktops

The Virtual Desktops workload gets acceptable performance for up to 2 clients.The detailed metrics from our trace playback are available here

Multi-Client NFS Performance for Professional Workloads

We looked at the multi-client performance of CIFS shares for professional workloads in the previous section. In a similar manner, we also evaluated the multi-client NFS performance of the Netgear RN202. Instead of the Windows 7 VMs, we used CentOS 7 VMs. The network configuration remained the same. The NFS share exported on the NAS was mounted with the following options.

<NAS_IP>:/PATH_TO_NFS_SHARE /PATH_TO_LOCAL_MOUNT_FOLDER nfs rw,relatime,vers=3,rsize=32768,wsize=32768,namlen=255,hard,proto=tcp,timeo=600,retrans=2, sec=sys,mountaddr <NAS_IP>,mountvers=3,mountproto=udp,local_lock=none,addr=<NAS_IP> 0 0

Using a popular filer benchmarking program, we did played back the same multi-client real-world professional workload access traces used in the previous section. Similar to the strategy for the CIFS performance evaluation, failing to meet the required op rate criteria at a particular load point made us stop the testing a couple of load points down the road. The Netgear ReadyNAS RN202 with two 7200 RPM hard drives in RAID-1 can support recording of 10 or more video streams, but, performance for other workload traces was found wanting.

Database Operations

The Database Operations workload doesn't get acceptable performance even for one client. The detailed metrics from our trace playback are available here

Software Builds

The Software Builds workload also doesn't get acceptable performance for even one client. The detailed metrics from our trace playback are available here

Video Recording

The Video Recording workload gets acceptable performance for up to 10 clients (the maximum we tested).The detailed metrics from our trace playback are available here

Virtual Desktops

The Virtual Desktops workload also fails to get acceptable performance for even one client. The detailed metrics from our trace playback are available here

Miscellaneous Aspects and Final Words

In order to keep testing consistent across all 2-bay units, we performed all our expansion / rebuild testing as well as power consumption evaluation with the unit configured in RAID-1. The disks used for benchmarking (Western Digital WD4000FYYZ) were also used in this section. The table below presents the average power consumption of the unit as well as time taken for various RAID-related activities.

Netgear RN202 RAID Expansion and Rebuild / Power Consumption
Activity	Duration (HH:MM:SS)	Avg. Power (W)
Single Disk Init	-	18.55 W
JBOD to RAID-1 Migration	08:45:42	29.33 W
RAID-1 Rebuild	08:42:18	29.37 W

The graphs below show the power consumption and rebuild duration when repairing a RAID-1 volume for the various 2-bay NAS units that have been evaluated before.

The Netgear RN202 is not the fastest or most power efficient system when it comes to RAID rebuilt, but it is in the top half of the set of 2-bay units that we have evaluated so far.

Concluding Remarks

The Annapurna Labs ARM Cortex A15 platform provides more CPU power and has the appropriate hardware acceleration blocks for the NAS market. With only two bays, the unit is more suited for home consumers. Even though link aggregatin is supported, the platform can't saturate the bonded link (topping out around 170 MBps in the real life workloads that we tested). However, in the space where ARM-based units tend to not support 802.3ad link aggregation, the RN202 is definitely better placed.

In our opinion, the Netgear ReadyNAS RN202 and ReadyNAS OS stand out in the crowded NAS market because of the btrfs support and the set of business features that enable interesting use-cases in the consumer market. Netgear's extensive networking background also allows for seamless VPN and cloud integration in ReadyNAS OS 6.

One of the interesting business features that Netgear has enabled in the RN202 is 'ReadyNAS Replicate'. This allows two ReadyNAS units connected to the Internet (and associated with the same ReadyCLOUD account) to be able to back up to each other via a secure peer-to-peer connection. This feature enables easy configuration of off-site backups. With certain tweaks (such as ability to do backups to a ReadyNAS associated with a different account), this feature could do well in the consumer market.

On the whole, ReadyNAS OS has slightly more than the bare minimum feature set required of a modern commercial off-the-shelf NAS operating system. The ReadyNAS RN202's hardware feature set makes is a good candidate for home consumers and SOHO installations. Taken in the context of units such as the QNAP TS-231 (with two network links and a sub-$200 price tag), the $282 diskless pricing of the RN202 is a bit too high for our liking.