It's a RAID!

The magazine of the Melbourne PC User Group
It's a RAID!

RAID may not be something you use or have considered, but it's certainly worth reading up on. Wikipedia is a good starting place.

New RAID levels, technologies and interfaces make choosing a RAID controller more than just a choice between price and performance. These tips from manufacturer Adaptec provide insight to help ensure that you get exactly the right controller to suit your specific data protection needs.

1. Choose the correct bus interface for your needs — forward- or backward-compatibility

Choosing a RAID controller currently means choosing between a PCIe bus interface and a PCI-X interface; and it's important to make the selection that will suit your needs. PCI-X has the advantage of being backwardly compatible with the older PCI interface. But it's parallel and half-duplex bidirectional, and the bus runs only as fast as the slowest device. PCIe is the new forward-looking standard, and is intended to cope with the performance and scalability demands for at least the next decade.

PCIe has the advantages of being serial, full- duplex bidirectional, and devices are able to independently negotiate the bus speed.

2. Find an easy-to-use management interface

A complex interface will make it harder for you to get the protection you want, be troublesome to set up initially, and more difficult to use if you do need to recover your data. In addition, having as many different interfaces as controllers you purchase can make life even more confusing. Make protecting your data easy by using controllers with one simple interface.

3. Which RAID level?

Choosing which RAID level you need is getting increasingly complex, as new RAID levels are introduced. Ensuring that you have the correct level for your needs is vital; get an expert opinion about which RAID levels you will get with your controller. To choose the RAID level that's right for you, begin by considering the factors below. Each of these factors becomes a trade-off for another.

Cost of disk storage
Data protection or data availability required (low, medium or high)
Performance requirements (low, medium or high)

4. RAID level migration

What if your data is already protected by RAID, but your requirements have since changed, and you'd now benefit from a different RAID level? If you have to move it to a new server, is it worth updating your RAID level at the same time? This is possible, and with some RAID controllers straightforward. So consider how complex the process is to migrate your data from your current RAID to your new one and consider whether RAID level migration is something you need.

The basic idea of RAID (Redundant Array of Independent Disks) is to combine multiple inexpensive disk drives into an array of disk drives to obtain performance. capacity and reliability that exceeds that of a single large drive. The array of drives appears to the host computer as a single logical drive.

The MeanTime Between Failure (MTBF) of the array is equal to the MTBF of an individual drive. divided by the number of drives in the array. Because of this, the MTBF of a non-redundant array (RAID 0) is too low for mission-critical systems. However, disk arrays can be made fault-tolerant by redundantly storing information in various ways.

Five types of array architectures. RAID 1 through RAID 5, were originally defined, each provides disk fault-
tolerance with different compromises in features and performance. In addition to these five redundant array architectures, it has become popular to refer to a non-redundant array of disk drives as a RAID 0 array.

5. How much more data capacity will you need?

When calculating the size of your new server, do you base your data requirements on your existing volume of data, or try to estimate how much storage space you'll need by the time you retire your server? One way of getting around the issue of buying all your future requirements now, is expanding the capacity as you go. This can be done by adding larger disks to your server, but different RAID controllers will deal with additional capacity in different ways, some a lot simpler and smoother than others, so be sure that you know how your potential solutions will deal with this.

Compare the Types of RAID Implementation
Type	Description	Advantages	Ideal for
Software based	Included in the Operating System (OS) such as Windows, Netware and Linux. All RAID functions are handled by the host CPU which can severely tax its ability to perform other computations.	Low price Only requires a standard HBA	Best used for large block applications such as data warehousing or video streaming. Also where servers have the available CPU cycles to manage the I/O intensive operations certain RAID levels require.
Hardware based	Processor-intensive RAID operations are off-loaded from the host CPU to enhance performance. Battery back write back cache can dramatically increase performance while maintaining risk of data loss.	Data protection and performance benefits of RAID More robust fault tolerance features and increased performance vs. software based RAID	Best used for small block applications such as transaction orientated databases and web servers.
External hardware	Connects to the server via a standard HBA or controller. RAID functions are performed on a microprocessor located on the external RAID controller independent of the host.	OS independent	Build high quality storage systems for high-end servers.

6. A limitation of SATA

If you're looking at expanding the capacity of your server with JBODs, there's a more straightforward infrastructure than SATA - even if you want to use SATA disks. The SATA infrastructure allows for Port Multipliers, but they have drawbacks - they can't be daisy-chained, limiting their flexibility and expandability. They only support one active host connection at a time, significantly degrading effective throughput and allowing for potential misidentification of drives. SATA drives can be used in Serial Attached SCSI (SAS) infrastructures - the infrastructure that allows for up to 128 drives - and with one RAID controller, the cost-per drive of the controller is significantly lower.

7. Getting locked in to your operating system

Operating Systems are constantly being upgraded, and the original reasons for choosing one OS over another may not be the same in a year's time. Additionally, you may have different OSes on different systems today. Your choice of RAID controller shouldn't restrict either your choice of OS, or restrict the speed at which you can obtain an OS upgrade.

8. Reliability of the RAID code

One factor in your choice of RAID controller may be to consider of the reliability of the RAID code. Is it used in many other servers? Has it been around for long and thoroughly tested in the field? If you need to rely on the data on your server for seven years, has the RAID code that's storing it been around that long?

9. Hardware or software RAID?

The difference between hardware and software RAID isn't just the price. With hardware RAID, the calculations are carried out by the RAID controller, with software RAID they take place on the server's CPU. So, if the RAID calculations are fairly simple, say RAID 1 or RAID 10, and the server is fairly powerful, using software RAID shouldn't be much of a problem. But with more complex RAID level calculations (RAID 5EE or RAID 6 for example), using hardware RAID can be beneficial because the
RAID performance isn't compromised by the server's workload, nor are applications on the server compromised by the RAID workload.

With hardware RAID, the RAID functionality is also independent of the OS, and the simple HBA drivers required for a hardware RAID controller are usually available as part of the OS distribution. Also, if it has a battery, hardware RAID can run in write-back mode, adding another level of data protection.

10. Ensure the products you purchase have the support you need

Your controller should work straight out of the box. But what if your combined system has problems? Who do you turn to? Some controller manufacturers offer technical help - even if the issue isn't obviously with the controller.

Learn The Levels Of RAID
RAID Level	Description	Maximum # of Drives	Benefit
RAID 0 RAID 1 RAID 1E RAID 5 RAID 5EE RAID 6 RAID 10 RAID 50 RAID 60	Data striping (no data protection) Disk monitoring Disk monitoring Data striping with distributed parity Data striping with distributed parity with hotspare integrated into the array Data striping with dual distributed parity Data striping of RAID 1 arrays Data striping of RAID 5 arrays Data striping of RAID 6 arrays	2 2 3 3 4 4 4 6 8	Highest performance Highest data protection Highest data protection for an odd number of disks Best cost/performance for multi-drive environments The cost/performance of RAID 5 without setting aside a dedicated hotspare disk Highest fault tolerance with the ability to survive two disk failures Highest performance with highest data protection Increased capacity and performance for multi-array RAID 5 environments Highest fault tolerance with highest data protection

Reprinted from the September 2008 issue of PC Update, the magazine of Melbourne PC User Group, Australia

[ About Melbourne PC User Group ]