Tag: RAID

Article – From the horse’s mouth

QNAP

TBS-453A NASBook Ultra-Compact SSD NAS

Product Page

My Comments

QNAP has just lately offered the TBS-453A NASBook which is an ultra-compact 4-drive NAS that is designed to work with solid-state drives rather than hard disks,

But this solid-state-drive NAS is a different breed to the portable NAS units offered to mobile users who exchange data between laptops and tablets. This is about Ethernet rather than Wi-Fi connectivity along with a 4-disk RAID array for performance or data safety, and increased functionality thanks to a desktop-grade NAS app store offered by QNAP.

The drives that are preferred for this device are the M.2 type that are typically used for Ultrabooks and 2-in-1s with this device being about the size of a B5 notebook. As well, it is powered by a 19-volt power brick but can accept voltages between 10 volts and 20 volts DC. This makes it suitable for a wide range if industrial and similar uses and could appeal to automotive and marine use, if there was a way to support externally-switched power control expected for such use.

It also has a 4-port Ethernet switch for one network along with a single Gigabit Ethernet socket for another network and this can be set up to work effectively like a router or to serve its own network.

As well, like other QNAP NAS units, this implements the QTS operating system with an increasing array of apps for business and personal use. But it also can be set up as a Linux computer by implementing a virtualised dual-operating-system setup and has the necessary “console” connections for this purpose. That is HDMI for display and audio connection along with USB connections for keyboard and mouse purposes. Like some recent QNAP NAS units, it also has its own audio interface which makes it appeal as a multimedia computer in its own right.

You can also expand this book-size NAS to work with a regular disk array by installing UX-800P and UX-500P multi-disk expansion modules this allowing you to create RAID disk arrays with these disks.

Tool or toy?

This kind of “far-fetched” cutting-edge network-attached-storage device could be easily dismissed as a toy rather than a tool, but there are some applications where it could earn its keep as a tool.

One would be to exist as a highly-capable Ethernet-connected portable NAS or server device that you use in a “mobile-office” setting. Similarly, you could see this being used with a home network where you want the multimedia functionality like DLNA Network Media Server functionality looked after but without dealing with a noisy or power-inefficient device. This would also win favours with home-AV manufacturers and distributors who are showcasing their network-capable hi-fi and home-theatre equipment at the hotel-based hi-fi shows like what the Chester Group are running. As well, QNAP are pitching this device in the Internet Of Things and building-security applications scenarios where this NAS could record data from sensor-type devices or visual images from network cameras yet be able to work with low power demands.

QNAP could see the TBS-453A as an effort to approach vehicle and marine applications for NAS devices along with courting mobile workgroups, remote data collection or other setups where portability, low power consumption or reduced operating noise are highly valued.

WD MyCloud EX2 2-disk NAS – has a 2-disk RAID setup

All network-attached-storage units that have two or more drive bays in them offer different ways to make use of the hard disks installed in these drive bays. These are primarily about creating one logical disk volume out of the many disk drives.

You may also find multiple-disk arrays being implemented in so-called “Direct Attached Storage” devices which connect to your computer as if they are a peripheral or are integrated in the computer. These are typically used for computer setups where read-write performance for secondary storage is considered important like video editing or for computers that work as servers.

WD MyCloud EX4 NAS – can be set up as a 4-disk RAID array

The most common setups are described as “RAID” or “Redundant Array Of Independent Disks”. These setups gang multiple hard disks (or solid-state drives) to improve data throughput, effective disk capacity or system fault-tolerance.

Multi-Drive Disk setups

RAID setups

It is important to remember that a RAID setup that is about fault tolerance doesn’t obviate the need to back up the contents of a NAS. This is something you can do with a USB hard disk connected to the NAS, another NAS on the same network or connected via the Internet or an online storage or backup service.

RAID 0

RAID 0 Data Striping across disks

Here, this creates one logical volume with the data spread across the disks, a method known as “striping”. Each block of data is sequentially stored across each physical disk rather than a disk being filled with data then another disk being subsequently filled with data.

This allows for increased capacity and read / write data throughput, but loses on fault tolerance because the disk array is no good and the data is lost if one of the drives fails.

Volume Capacity: Number of Disks x Size of smallest disk

RAID 1

RAID 1 – Disk Mirroring

This setup creates a logical volume with the data duplicated on each physical drive. a method known as “mirroring”.

The main advantage here is increased fault-tolerance because if a disk dies, you still have access to the the data on the other disk. There is also another advantage of increased read throughput because both physical disks can be read at the same time.

The only limitations here are the volume capacity which is the size of the smallest disk in the array along with the write speed because the disk controller has to write the same data to multiple disks. It is infact a preferred RAID array setup for a 2-bay NAS due to the fault-tolerance.

Volume Capacity= Size of smallest disk in the bunch

RAID 5

RAID 5 Data Striping with a parity block

This setup works between data capacity and fault tolerance in a very interesting way. It is because the RAID 5 setup creates “parity” data. This is used in computing as a fault-tolerance measure because an algorithm can use this data along with the “known-to-be-good” data to reconstitute data lost in transmission.

Here, a RAID 5 array stripes data across the physical disk collection but inserts a block of parity data at regular intervals as part of this “striping” so as to create some form of fault-tolerance. Then the RAID 5 disk controller reconstitutes data from parity and available “known-to-be-good” data if things start to go wrong with a disk.

The advantages in these setups are the disk capacity, the read throughput and the fault tolerance but there is a performance reduction for those systems that do a lot of data writing.

Volume Capacity: (Number of disks – 1) x smallest disk size

RAID 6

RAID 6 Data striping with two-block parity

RAID 6 works in a similar manner to RAID 5 in that it stripes data across multiple physical disks and creates a parity block for fault-tolerance. But a RAID 6 array creates another parity block to increase the amount of fault tolerance in the setup.

Volume capacity: (Number of disks-2) x smallest disk size

RAID 10 (1+0)

RAID 10 A combination of data striping and disk mirroring

A setup that is used with 4-disk RAID arrays is the RAID 10 array also known as the RAID 1+0 array which is a combination of both the RAID 1 setup and the RAID 0 setup.

Here, there are two collections of disks with one collection keeping copies of the data held on the other collection. Each collection has its data “striped” across the disks for capacity and performance.

The core benefit with a RAID 10 setup is that there is increased write throughput which can come in handy with write-intensive setups like databases. This is in addition to the fault tolerance provided by mirroring along with the read performance provided by striping.

Volume Capacity: Combined size of two of the smallest disks

Non-RAID setups

JBOD – Disks as separate volumes

JBOD

This setup, known as “Just a Bunch Of Disks” is simply about each physical disk being treated by the NAS as a separate logical volume. It can be useful if you want to maintain separate data on each disk under a separate volume name.

Spanning

Disk Spanning, sometimes known as JBOD by some manufacturers

The “spanning” setup simply is based on data filling up one disk then filling up another disk in that same volume.

Different options available

Automatic RAID setups

The NETGEAR ReadyNAS on the right can implement X-RAID automatic RAID setup

An increasing number of manufacturers use an “automatic RAID” setup like Synology’s Hybrid RAID or NETGEAR’s X-RAID. These are RAID setups that are optimised to mix different-sized hard disks so that these arrays work to maximise useable capacity, disk performance and fault-tolerance.

Manufacturers pitch these RAID setups for people new to NAS or disk-array management who are thinking about how much redundant storage is needed to balance capacity and fault tolerance. They also encourage the customers to “build out” a RAID array as and when they can afford the extra disks.

Hot-spare disks

Thecus N5810PRO small business NAS is able to implement a hot-spare disk for high RAID availability

Another feature offered mainly with small-business NAS units is the addition of a hot-spare disk. Such RAID arrays will have a separate hard disk that isn’t used unless one of the disks in that array fails.  These setups are preferred for environments where there is emphasis on a multi-disk array that is to be highly available at peak performance.

Hot-swap setups

An increasing number of prosumer and small-business NAS units come with a “hot-swap” functionality where you can swap out the hard disk while the NAS is in operation. This is more so for replacing faulty disks that are degrading a RAID array’s performance and is more relevant with RAID 1, RAID 5, RAID 10 and “automatic RAID” setups.

Upsizing a NAS’s RAID array

QNAP 2-disk NAS – capable of setting up a highly-available high-performance RAID1 array

Upsizing a RAID array is something you could be tempted to do, especially as hard-disk prices gradually become cheaper and the time when one hard disk in a RAID array fails may be the time to upsize it.

But this can be difficult. Here, you would need to copy out all of the data to storage with the same volume capacity as your NAS’s current RAID array. Then you would have to simultaneously replace the disks in that array with units of the same but higher capacity before copying back the data. This may be easy to achieve with a 2-bay NAS.

Or you could migrate a 2-disk RAID 1 array in a 4-bay NAS to a RAID 5 array while adding a higher-capacity disk to that array. Here, you get increased capacity on the new disk due to the smaller disks being combined for real data use while space on the larger disk is allocated for parity data. Then you would need to swap out the small disks in that array with the larger disks as a way to gradually increase the volume’s useful size.

The automatic RAID setups make it easier to upsize your NAS as you can afford it and manage the right amount of redundant storage needed for your data.

The best RAID array setup for your needs and your NAS

The RAID array that you set your NAS up with depends on the number of drive bays the device has along with the number of disks you have. But these suggests are based on setups that are cost-effective yet yield high availability . They would also yield high read performance especially for multimedia applications. It is also a good idea to populate your multiple-bay NAS with drives of the same capacity when you are setting a new unit up.

A 2-bay NAS would be best set up as a RAID-1 array in order to implement the mirroring ability for high availability and increased read throughput which is necessary for video files streamed using DLNA.

A 4-bay NAS would be best set up as a RAID-5 array of at least three disks of the same size. There is the ability to make use of the capacity yet use the parity blocks to keep the data available should one of the disks keel over.

Conclusion

Once you understand how the various RAID and other multi-disk arrays work, you can choose the most cost-effective way to have your data stored for capacity, performance and high availability with your NAS.

On Tuesday 29 July, I had attended a Western Digital reseller presentation where WD were premiering their latest additions to the Red Series hard-disk range for NAS units. These are the WD Red Pro series that are pitched at heavy-duty applications centred around the many-bay units that can be mounted in a 19″ standard equipment rack as well as variants of the WD Red range that have 5 Terabyte and 6 Terabyte capacities.

The increasing relevance of the network-attached storage unit

D-Link DNS-320L 2-bay NAS

A device that is appearing in more home and small-business networks is the network-attached-storage unit. This is a dedicated unit that shares data held on at least one hard disk across a network.

But increasingly these devices are being able to do more than this due to the vendors marketing their NAS units as a “platform” with a plethora of apps developed or ported by the vendor for these devices. This is augmented by an increasing number of manufacturers who are integrating the kind of processors used in regular-computing or enterprise-grade server applications in these devices, some of which you could describe as being like a compact desktop PC.

They are increasingly relevant in the small-business scene where they can serve as a backup location or central storage for that business’ computers. An increasing number of these units can implement “virtualization” where they can work as one or more different computer systems. As well, platform-based NAS units offer applications like video-surveillance recording, digital signage, and enterprise-grade “advanced-storage” setups like iSCSI or SAN. For that matter, some of the high-end desktop NAS units can be purposed as branch-level “on-ramps” for a full-blown enterprise-wide computing setup.

ASUSTor AS-204TE 4-bay NAS – “Data central” for a small business

A few creative-skills professionals appeared at this presentation to demonstrate how the NAS fits in to their trade. In most of these cases, these users store only the data they need to work with at a given moment on their iMac’s main hard disk and keep the rest of the data on a NAS. As well, the units even serve as central content libraries for raw material or finished projects. This appeals very strongly to multi-person projects like film and video work where version-control is important.

But, in the home, they are appealing as systems to hold your audio, video or image content and make it available to network-capable AV devices. As well, most NAS vendors are pitching these systems as a “personal cloud” that is an alternative to services like Dropbox or OneDrive. As I have mentioned before, vendors are offering DVR abilities so that customers can connect the NAS to a USB digital-TV tuner module or broadcast-LAN unit to make it become the equivalent of a TiVo.

Making the NAS appeal to “Average Joe”

QNAP TS-251 2-bay NAS

One of WD’s representatives found that there is a problem with selling a NAS in a “big-box” store like Harvey Norman or JB Hi-Fi. Here it is about identifying the value that these devices have for the average “Joe in the suburbs” who is content with using a USB external hard disk as a backup or offload tool for their home computer. Typically the home network is implemented by these users just to facilitate Internet access and, perhaps, share a printer.

What needs to happen to make the NAS appeal to “Joe in the suburbs” is that a NAS makes more sense as an always-available content library or data store, especially if you have or intend to buy another computer, a mobile device like an iPad or network-capable AV equipment including most recent games consoles or smart TVs. A good question to address is the number of digital pictures you take or hours of digital video footage you make and the number of CDs you rip or digital music files you buy from iTunes and similar services and how you can make them available around the home.

As well, one or more legitimate “download-to-own” video-content services that can allow you to store your movies that you downloaded to a NAS can legitimize the value prospect of these devices to “Average Joe”.

WD internal hard-disk lineup and the RED Series

What has happened over the last few years is that WD have re-factored their regular-duty computer hard disk lineup in to distinct ranges denoted by colour as shown below:

This kind of product lineup avoids the practice where most user-installable desktop hard disks are sold to users as a “jack of all trades” basis without awareness of disks that are optimized for particular data-storage needs. For example, a person who is running that “ultimate gaming rig” to impress others at the LAN party would be after something that is about performance whereas a NAS or server user is after something that is about consistent reliable operation for something that is always available.

What are the WD Red Series hard disks and what makes them special

One of the many business-class “pizza box” NAS units that works with the WD Red Pro hard disk

WD were the first company to develop and launch a hard disk that is optimized for the operating conditions that a network-attached storage device will throw at it. Previously, a NAS used regular desktop hard disks as its storage and these disks were seen more as a “jack of all trades, master of none” when it came to network storage requirements.

The key features for this range include:

• Compatibility with the different operating conditions that different vendors’ NAS units will throw at the system. This includes dealing with different power-supply conditions, the hardware interfaces used in the NAS units or how they present to the software that is used in these devices.
• Always-on reliability. The typical network-attached storage system is expected to be on all the time, ready to serve data when needed and is often seen as being “Data Central” for the home or business network. Here, these hard disks are expected to be spinning. It includes the provision of NASAware firmware on the hard disks to deal with situations like power loss or power disruption that can affect system reliability.
• RAID-friendly design. WD have factored in vibration-control measures in order to cope with the typical multi-bay RAID-capable NAS. This is because with many hard disks in close physical proximity to each other, there is increased vibration when the NAS is moving data to multiple disks at the same time such as “mirroring” data across multiple disks. The RED series implement software or hardware measures to counteract the effects of continued vibration that occurs in these setups.
  As well this design also is supported with hard-disk firmware that can assure proper error recovery in the many-disk RAID arrays used in these devices thus avoiding the risk of underperforming RAID setups.
• Power flexibility and efficiency. The WD RED series of NAS hard drives are optimized for varying power conditions that can be thrown at them, such as when a multi-bay NAS is being started or for different NAS units that have different power-supply characteristics. This also includes being designed for power efficiency in an always-on environment, even though most recent desktop NAS units implement on-demand “spin-up / spin-down” measures to save energy.

WD MyCloud EX Series NAS units able to benefit from the 6Tb WD Red

The newly-released 5Tb and 6Tb capacities appeal to all NAS designs in a lot of ways. For example, you could set up a 6Tb single-disk NAS or use two of the 6Tb hard disks in a dual-disk NAS configured for RAID 1 to have a fail-safe 6Tb data volume that can also handle higher data throughputs. You could even run up to 24Tb in a four-bay NAS or 30Tb in a five-bay NAS, including implementing various RAID data-replication setups for fail-safe or high-throughput operation.

Even the way the hard drives are designed have an efficiency and density advantage over the competition. For example, the 6Tb drives maintain 5 platters with 1.2Tb per platter rather than 6 platters with 1Tb per platter. This means that there isn’t much mechanical effort needed on the spindle motor to spin up the disk. As well, the drive housing can fit in to most NAS drive bays without being unnecessarily stout. They also maintain a 64Mb local hardware cache for improved operation efficiency.

The new WD Red Pro lineup

This lineup of NAS hard disks is optimized for the rack-mount large-business-class NAS system and is built towards higher performance and reliability in these many-bay systems. These would be able to handle a greater workload, which would be representative of a larger high-traffic business. Some people have put forward questions about using one of these hard disks in a small desktop NAS but it wasn’t found to be worth it for the kind of use that this class of NAS would typically be put to. But on the other hand, I would see them as being of use with the smaller units that serve branch-based “on-ramp” applications for enterprise data infrastructures.

Using the WD Red or the WD Purple disks for video-surveillance applications

QNAP TS-EC880U-RP business-class “pizza box” NAS that works with the WD Red Pro hard disk

Some questions were raised about implementing WD Purple hard disks in a regular NAS that was running one of the video-surveillance apps offered by the vendor as part of their application platform. The WD presenters recommended that the WD Purple disks go in dedicated DVR equipment that is optimized for the task rather than NAS units running these platform apps. Instead, they recommended the use of WD Red disks in these “NAS+software” setups, more likely because the NAS may be tasked to do other network-storage activities like being “Data Central”.

Can my NAS handle 6-Terabyte disks

A situation that one can easily run into with any computing equipment is that the equipment’s operating system or firmware can impose an arbitrary limit on the size of storage media. Here, if you supply storage media that is greater than this maximum allowed in this software, the software could throw up errors or simply fail because it can’t address all of the storage media’s useable capacity. This problem shows up when storage-media manufacturers release higher-capacity media after the software was “set in stone”.

For example, the older versions of MS-DOS and some other desktop operating systems couldn’t handle large capacity hard disks as a single logical volume. So computer users had to partition larger-capacity hard disks in to multiple logical volumes in order to make use of this space. As well, I had used an older digital camera that worked with SmartMemory cards and couldn’t use newer higher capacities of these cards. Here, I had to look around for cards of a particular capacity to keep as “spare film” for the camera.

Most of the NAS platforms can support this capacity out of the box or may require you to wait on an interim update for the new capacities to be supported. WD have provided a compatibility list which allows you to find what of the WD Red range can be supported by your NAS box. This includes issues like maximum capacities that these systems have. It is also worth checking on the vendor’s Web site for newer or impending software updates.

Conclusion

If you are thinking of buying an enclosure-only NAS or “upsizing” your existing NAS, you can head towards the newer 5Tb or 6Tb disks that WD offers for increased capacity. As well, your heavy-duty many-bay business-grade NAS can be treated to the WD Red Pro disks that are appropriate to its usage nature and performance level.

An idea that has come to me for small businesses that have a handful of staff is the concept of a small server that isn’t too overpowered for their needs but allows them to “grow up”.  What I mean by that is that  a small business that runs with a few regular desktop or laptop computers but without a server may not be seen by some as being a “grown-up” business as far as their IT needs are concerned. This may be due to absence of flexibility in these setups or not being able to cope with larger volumes of business due to the smaller size of these setups.

Here I am reviewing the Western Digital Sentinel DS5100 NAS which is a four-disk NAS that runs on the Essentials version of the Windows Server 2012 operating system rather than a shoehorned version of Linux which most NAS units run on. This software is licensed and optimised for 25 users and 50 computers.

The Sentinel DS6100 series comes with increased capacity and two power supplies which you can set up for increased power-supply redundancy and reliability. It also runs with two 2.5” boot drives rather than one boot drive also for increased reliability.

Price

WD Sentinel DS5100

WD Sentinel DS6100

Specifications

The Network-Attached Storage System itself

Small size alongside a loaf of bread

This is a 4-disk NAS that runs with Intel Xeon E3-1220L horsepower but is in a cabinet half the size of a breadbox. It is well-built and hasn’t shown any symptoms of excessive vibration or noise, although the large fan does spin up hard during the initial boot-up phase. Even transferring data to a parity-arranged RAID setup where two or more disks would be spun up meant that the system was relatively quiet.

This is because the power supply is a laptop-style “lump” that is outside the NAS’s case rather than physically integrated in the unit. This allows for quieter operation and the smaller size that this unit has.

The WD Sentinel DS Series of NAS units – effectively a “business in a box”

Setup and use

Connections – 2 Gigabit Ethernet, 4 USB 3.0, 2 USB 2.0, VGA, power

The setup routine requires that you to use a dedicated keyboard, monitor with VGA connection and mouse as a “console” to get Windows Server 2012 Essentials going on the WD Sentinel. After that, you have to go to the “//servername/connect” URL to download connection software for your operating system so you can manage the WD Sentinel from your computer.

The Windows Server 2012 R2 Essentials software can support up to 25 users on 50 machines initially out-of-the-box but there is the ability for the small business to upgrade to Windows Server 2012 Standard. The fact that this box runs on Windows Server means that it could run as a server for a variety of business-grade software applications such as being part of a “point-of-sale” system.

The throughput was pretty good enough for most light-duty file-by-file work, being able to accept 220Gb over two hours. This could satisfy most small-business data requirements especially at the early stages.

Limitations and Points Of Improvement

Access door to hard drives

For a person who isn’t used to Windows Server, it may appear to be very daunting to do tasks like setting up the data volumes. WD could improve this with a wizard that simplifies and directs setup towards a RAID-5 setup. Similarly, the Windows Server Essentials misses out on the ability to run the Exchange Server and the main file / Web / application server function in one box which may put a limit on its use as the dream “business in a box” server.

Microsoft could pitch a lightweight variant of Exchange Server for use on Windows Server Essentials setups in order to cater for on-premises setups with these “business-in-a-box” servers and could work towards simplifying Windows Server Essentials setup for the businesses who don’t have ready access to IT staff. As well, small business needs to be aware of line-of-business software that works with Windows Server Essentials in an easy-to-manage manner and can run on these servers.

This class of server could be a chance for WD, Seagate and others, along with Microsoft to achieve the goal of the small “business-in-a-box” server that is highly capable. For example, more-powerful highly-compact “breadbox” servers could appear, with some being able to run from a vehicle’s or boat’s battery leading towards the goal of server-assisted big-time computing for businesses that work out of a vehicle or boat.

As well, a computer dealer could offer a “get-you-going” pack with one of these servers along with a UPS at a price that could please small business so they can have that “foot-in-the-door” with server-based computing.

Conclusion

Personally, I would pitch the WD Sentinell DS5100 or DS6100 for one of many applications. For example, a person who is teaching or learning Windows Server skills could use this box rather than a surplus desktop “tower” as part of this activity. Similarly, this could be set up as an entry-level “business in a box” server for a small shop moving from the cash register or single-terminal POS setup to a more flexible setup.

Similarly, this box, with Windows Server’s BranchCache feature could work well with small businesses who are running from a few different locations or want to establish a transportable location like the so-called “pop-up” branch. Here, it means that the unit could serve as a “local cache” server for these remote or transportable locations thus mitigating poor-quality or expensive Internet backlinks.