RAID or Redundant Array of Inexpensive (or Independent) Disks, can improve disk drive performance by spreading data across multiple disks that are treated as one logical drive. A RAID subsystem can enhance system performance, provide fault tolerance, simplify the process of adding disk capacity, and make building extremely large disk volumes possible.
RAID has been around on UNIX and mainframe systems for many years, and the technology was built into Windows NT from the start. With the flurry of activity surrounding NT clusters and server scalability, RAID has recently gained new exposure in the NT market (see the June 1997 issue for more information about NT-based cluster solutions). But what is RAID? And how can it help you improve your NT systems' performance and reliability?
Let's explore the answers to these questions with a detailed technical look at RAID options for NT systems. I discuss the best RAID levels to use for optimizing performance and fault tolerance, and provide some general guidelines for choosing a RAID system. For an introduction to RAID, see Raid Levels
Hardware and Software RAID
The two types of RAID are hardware RAID, in which the disk controller
performs the RAID functions, and software RAID, in which the operating system
performs RAID functions. NT 4.0 lets you use hardware- or software-based
solutions or combine the two to achieve the best performance and fault
tolerance.
| Enhanced Disk Storage for Windows NT RAID distributes data on multiple disks, boosting server performance, protecting data--or both. |
Many vendors, including Adaptec, American Megatrends (AMI), Compaq, and Mylex, provide hardware RAID solutions (disk controllers and array chassis) that offer many of the RAID levels listed in "RAID Levels." RAID 0, 1, and 5 are the most common. As a rule, hardware-based RAID solutions are faster and more reliable than software-based ones. They also offer a greater range of configuration options. Of course, they're more expensive than using NT's built-in RAID, but if you want the best performance, strongly consider including hardware-based RAID in your overall system budget.
NT supports RAID functionality, offering software settings for RAID 0 in NT Workstation and 0, 1, and 5 in NT Server. The advantages of software RAID are the convenience of built-in software and cost. However, performing RAID functions through the operating system instead of offloading them to a separate controller can slow server performance.
As you can see in "RAID Levels," each level has different performance characteristics, fault-tolerance capabilities, and drive usages. Some levels offer excellent all-around performance, and others sacrifice this performance to gain fault tolerance (for more information about how the performance of RAID 0, 5, and 10 compare, see "Optimizing Exchange Server," November 1996). Because each level is suited to a particular environment, your main challenge when choosing a RAID subsystem is to decide which RAID level to use on your server under what conditions. Let's look at the tradeoffs of the RAID levels most commonly used in NT systems and some tips for choosing the best RAID for your system. (See "RAID Tips," for some RAID-optimization hints. And for information about other ways to improve disk subsystem performance besides RAID, see "Pumping Up Your Server")
Optimizing for Performance: RAID 0
Disk performance is a critical factor in server performance. Disk access is
much slower than memory access. Therefore, the faster your disk I/O, the faster
your server's response time. As a rule, RAID 0 (i.e., plain disk striping)
provides the fastest I/O and thus the best performance.
RAID 0, or normal striping, splits data blocks (chunks of data) across multiple disks simultaneously. The group of disk drives containing the split data is called a stripe set; the size of each data piece depends on how many disks are in the stripe set. Striping means all drives are active for every I/O transaction and that each drive in the stripe set does less work per transaction. Less work means faster performance.
You can immediately benefit from RAID 0 by using NT's Disk Administrator to create stripe sets. This approach lets you create larger disk volumes under NTFS (FAT has a 2GB partition limit) and improves disk I/O performance.
Software striping via Disk Administrator is useful for just about any application, but with some cautions. First, software striping causes some minimal additional CPU overhead because NT now has to calculate striping instead of just passing I/O requests to the disk controllers. However, with today's fast CPUs, this overhead is not a problem because the processing takes a very small percentage of the CPU's overall capacity and the performance benefit of using multiple drives is greater than the performance hit. Systems with old processors (386, 486, or even slow Pentiums) may have more difficulty, and you need to augment them with a hardware RAID controller, which offloads RAID calculations from the system's main CPU or CPUs.
Second, be careful of where disks are located in the system. If you stripe disks on two or more SCSI controllers (called controller multiplexing), you're asking NT to calculate which data goes where in addition to figuring out the striping, not to mention processing overhead, system bus traffic, and processor interrupts for handling multiple cards. Again, older systems may have trouble handling this processing.
Try to stripe disks only on the same controller for the best performance, unless the capacity simply isn't enough from one SCSI card. You can compensate for the above problems by using a hardware RAID controller that has specific circuitry for handling these calculations and multiple channels for enhancing performance and adding capacity (a multichannel card uses only one interrupt).
The big drawback of RAID 0 is that it offers no fault tolerance: If one drive in the stripe set dies, the entire volume is unrecoverable. Also, the number of drives you use in a stripe set has a point of diminishing returns.
For example, the results explained in "Microsoft SQL Server 6.5 Scaleability," (January 1997) showed that six drives were the effective limit for a Compaq ProLiant 5000 with a Smart 2/P Array Controller; more drives improved performance minimally. This minor improvement is because the mechanics of the situation catch up with you (as the number of drives in the stripe set goes up, the block size goes down; if the block size drops below the stripe width, the advantages of striping diminish). In addition, you saturate the SCSI channel with too many drives. New controllers with faster hardware, such as Wide SCSI-3 and Ultra-2, raise this limit because they can run at 40MBps or 80MBps and use wider (32-bit) data words. (See ">RAID-Related Terms," for definitions of the SCSI standards and other terms.) Another issue is that more drives mean greater probability for failure.
Some experts recommend that you never use RAID 0 alone on a server. However, the question is one of cost vs. performance, so RAID 0 with an aggressive backup policy may be worthwhile.
Optimizing for Fault Tolerance: RAID 1 and 5
Optimizing your server's disk storage is a balancing act: You want the best
possible performance, but you need to protect your data, too. RAID 1 and RAID 5
are two widely used methods for protecting data.
RAID 1, disk mirroring, is most often used for smaller critical data volumes. It gives you complete fault tolerance (either drive in the mirror set can fail without affecting system integrity or performance) and slightly better performance than no RAID. The tradeoff? Because both drives are exact copies of each other, you get only 50 percent of the disk capacity you purchased.
RAID 5 is the most commonly used option for fault-tolerant disk volumes in NT because most manufacturers implement and support this method, it is part of NT Server, and it offers a reasonable compromise between performance and disk capacity. RAID 5 offers enhanced performance, protection, and far less capacity loss than RAID 1. Because you can build a RAID 5 volume out of as few as three drives, the maximum capacity you lose is 33 percent; the more drives you add, the less total space you lose. RAID 5 offers better I/O read performance than no RAID at all and in some cases, is even better than RAID 0 (because of the striping algorithm used). The drawback of RAID 5 is that write performance suffers significantly because every I/O operation requires a parity calculation. This performance hit in software RAID 5 is high; you'll probably want to use a fast RAID controller to compensate for the overhead.
The advantages to RAID 5 are that you can build very large fault-tolerant disk volumes, and any drive in the stripe set can fail without damaging data. However, fault tolerance doesn't mean you won't suffer a little if a drive fails. When one drive disappears from the stripe set, either your system CPU or the RAID controller must compensate on the fly by using the remaining data and parity information to reconstruct the data for every I/O request. Depending on your system and controller, this reconstruction could mean as much as a 50 percent performance hit on that volume--but at least you're still running!
In NT, this recovery process is automatic (as it is on hardware controllers). NT also automatically rebuilds the volume when you replace the faulty drive. As soon as the system gets a new drive, it begins the background process of reconstructing the data on the new drive in the same way it handles I/O requests on the fly (this process can take several hours, depending on the volume/disk size). The process slows performance (more with software RAID than on an accelerated controller), but as soon as reconstruction is finished, system operations return to normal.
Also note that in software RAID 5, you often cannot break the set to add a new drive. Such behavior makes RAID 5 on NT not such a great option, and some experts never recommend this approach. In contrast, this issue does not arise with hardware RAID.
Other Fault-Tolerance Options
Two additional RAID fault-tolerance hardware options are RAID 3 and 4.
Although they're less common on NT systems than other options (and NT does not
support them), they offer fault tolerance through striping with parity data.
In addition to providing fault tolerance through RAID, some disk controllers have special features that ensure availability in the event of a disk crash. Some RAID arrays feature hot-swap drives: You can remove and insert disks without powering off the disk cage or even the specific slot.
A hot swapcapable array should never go down due to a drive failure (barring component death of the backplane, faulty power supplies, or similar problems). Systems without hot-swap drives require you to power down the system to replace a bad drive. In systems with hot-swap bays, the controller/software detects the new drive coming online and begins repairing the volume.
Another option is a hot-spare--a drive in the array that waits in standby mode. If any other drive in the array fails, the system automatically switches over to the hot-spare and begins rebuilding, without administrator intervention. When you replace the faulty drive, it becomes the new hot-spare. You can enable hot-spares through the controller's BIOS or management software.
The Best of Both Worlds
A few combined RAID levels (e.g., RAID 10, 30, or 50) offer both performance
and fault tolerance by using two forms of RAID on the same logical volume at the
same time. As you might expect, you pay more to have both capabilities. This
extra cost is because NT's Disk Administrator tool alone won't let you combine
RAID levels; to do this, you must combine a hardware RAID controller with NT's
RAID software functions.
One combined RAID level is RAID 10, also called mirrored stripe sets (i.e., a RAID 0 stripe set is mirrored to another stripe set). RAID 10 offers excellent gains in read and write performance in sequential and random transaction environments. In fact, it's the best overall performer of all RAID levels. The cost, as with mirroring, is that you lose 50 percent of your planned disk capacity. But, where simple mirroring (RAID 1) costs you only one drive per mirrored set, RAID 10 costs you as many drives as are in the RAID 0 stripe set (which can get expensive). Like RAID 1, RAID 10 makes a fault-tolerant volume with the performance advantages of striping and no performance hit in the event of a drive failure.
Another combination of RAID 0 and 1 is RAID 01, or striped mirror sets, which has similar characteristics to RAID 10. The main difference between RAID 10 and 01 is which RAID level the hardware controller handles and which the software handles. In RAID 10, for example, if the software handles the striping, the controller performs the mirroring; in RAID 01, vice versa
Not all RAID controllers support level 10 or 01. You'll need to check which RAID levels a controller supports before you buy it. However, you can make combined RAID by using hardware for the first part (RAID 0 striping or RAID 1 mirroring) and software for the second (the alternative mirror or stripe, respectively). This solution does not perform as well as using a RAID hardware controller that can handle both at the same time. But you can still build high-performance, fault-tolerant disk volumes without replacing an existing RAID controller.
Other RAID levels, such as 30 and 50, can also enhance performance and fault tolerance, depending on your applications. With them, you can build very large disk volumes out of commodity drives. However, these RAID levels are of limited use in most low- to midrange NT server situations, unless your goal is to experiment or achieve new and interesting disk configurations. RAID 50 is a good option on an enterprise-scale server where you are trying to build a 500GB or even 1000GB disk volume.
The Right RAID
With the variety of available RAID options, you can choose the right balance
of performance and fault tolerance for your site. Mixing hardware and software
RAID lets you build disk subsystems specifically tailored to your needs, such as
extremely large disk volumes or multiple-faulttolerant arrays. Whatever
RAID you consider, it's a disk technology you can't afford to be without.