Common ATA Configurations for Server Platforms

Although SCSI hard disks have traditionally been used in midrange and high-end server platforms, ATA-based hard disks continue to be popular in entry-level servers. The superior performance and easier configuration offered by SATA hard disks make them a suitable choice for both entry-level and many midrange server implementations. The following sections provide recommendations for using ATA drives in a server platform.

ATA/ATAPI Configurations for Server Platforms

If you decide to use PATA hard disks in an entry-level server, you need to configure your server as follows:

• Connect the hard disk to the primary PATA (40-pin) connector on the motherboard.

• Connect the CD or DVD optical drive used for program loading or for creating backups to the secondary PATA connector on the motherboard.

• Choose a hard disk with a 7200RPM spin rate and an 8MB cache for best performance.

• Use 80-wire UDMA cables on both drives to permit maximum speed operation.

• Be sure to install the appropriate UDMA drivers in the operating system.

Note that most commercial entry-level servers no longer use PATA hard disks, although ATAPI CD and DVD drives continue to be popular.

SATA Configurations for Server Platforms

As noted earlier in this chapter, SATA drives have largely replaced PATA hard disks in entry-level servers, and they have made inroads into the midrange server market as well.

To improve reliability and performance, you should consider using SATA hard disks that meet the following standards:

• 7200RPM or 10,000RPM spin rates; drives with 10,000RPM spin rates are recommended when access time is more important than high capacity (10,000RPM spin rate drives are currently limited to 150GB per drive, while 7200RPM drives are available in capacities up to 500GB)

• 8MB or larger cache; some 16MB cache drives are now available

• Support for command queuing (tagged or native) to improve throughput; the type of command queuing supported must also be supported by the SATA host adapter used by the system

• Designed for higher reliability; look at factors such as MTBF, thermal management, and warranty

Each SATA hard disk uses a separate host adapter.

ATA/SATA RAID Configurations for Server Platforms

RAID, which stands for redundant array of independent (or inexpensive) disks, was designed to improve the fault tolerance and performance of computer storage systems. RAID was first developed at the University of California at Berkeley in 1987, and it was designed so that a group of smaller, less expensive drives could be interconnected with special hardware and software to make them appear as a single, larger drive to the system. By using multiple drives to act as one drive, fault tolerance and performance could be increased.

Initially, RAID was conceived to simply enable all the individual drives in the array to work together as a single, larger drive with the combined storage space of all the individual drives added together. However, this actually reduced reliability and didn't do much for performance, either. For example, if you had four drives connected in an array acting as one drive, you would be four times as likely to experience a drive failure than if you used just a single, larger drive. To improve its reliability and performance, the Berkeley scientists proposed six levels (corresponding to different methods) of RAID. These levels provide varying emphasis on fault tolerance (reliability), storage capacity, or performanceor a combination of the three.

An organization called the RAID Advisory Board (RAB) was formed in July 1992 to standardize, classify, and educate on the subject of RAID. RAB has developed specifications for RAID, a conformance program for the various RAID levels, and a classification program for RAID hardware.

Currently, RAB defines seven standard RAID levels, RAID 06. RAID is typically implemented by a RAID controller board, although software-only implementations are possible (but not recommended). The levels are as follows:

• RAID Level 0: Striping File data is written simultaneously to multiple drives in the array, which act as a single, larger drive. This level offers high read/write performance but very low reliability. It requires a minimum of two drives.

• RAID Level 1: Mirroring Data written to one drive is duplicated on another, providing excellent fault tolerance (if one drive fails, the other is used and no data is lost) but no real increase in performance compared to using a single drive. It requires a minimum of two drives (and has the same capacity as one drive).

• RAID Level 2: Bit-level ECC Data is split 1 bit at a time across multiple drives, and error correction codes (ECCs) are written to other drives. It is intended for storage devices that do not incorporate ECC internally (all SCSI and ATA drives have internal ECC). It provides high data rates with good fault tolerance, but large numbers of drives are required, and no commercial RAID 2 controllers or drives without ECC are available on the market.

• RAID Level 3: Striped with parity This level combines RAID Level 0 striping with an additional drive that is used for parity information. This RAID level is really an adaptation of RAID Level 0 that sacrifices some capacity, for the same number of drives. However, it also achieves a high level of data integrity or fault tolerance because data can usually be rebuilt if one drive fails. It requires a minimum of three drives (two or more for data and one for parity).

• RAID Level 4: Blocked data with parity This level is similar to RAID 3, except data is written in larger blocks to the independent drives, offering faster read performance with larger files. It requires a minimum of three drives (two or more for data and one for parity).

• RAID Level 5: Blocked data with distributed parity This level is similar to RAID 4, but it offers improved performance by distributing the parity stripes over a series of hard drives. It requires a minimum of three drives (two or more for data and one for parity).

• RAID Level 6: Blocked data with double distributed parity This level is similarto RAID 5, except parity information is written twice, using two different parity schemes, to provide even better fault tolerance in case of multiple drive failures. It requires a minimum of four drives (two or more for data and two for parity).

• RAID Level 10 (RAID 1+ 0): Striping plus mirroring This level is a combination of RAID 1 (mirroring) and RAID 0 (striping). Provides higher performance than RAID 1. It requires a minimum of four drives (two for mirroring; two for striping).

Additional RAID levels exist that are not supported by RAB but that are instead custom implementations by specific companies.

Note

Note that a higher RAID level number doesn't necessarily mean increased performance or fault tolerance; the numbered order of the RAID levels is entirely arbitrary.

At one time, virtually all RAID controllers were SCSI based, meaning that they used SCSI drives. For a professional setup, SCSI RAID is definitely the best choice because it combines the advantages of RAID with the advantages of SCSIan interface that was already designed to support multiple drives. Now, however, ATA RAID controllers are available that allow for even less expensive RAID implementations. These ATA RAID controllers are typically used in single-user systems for performance rather than reliability increases.

Most ATA RAID implementations are much simpler than the professional SCSI RAID adapters used on network file servers. ATA RAID is designed more for an individual who is seeking performance or simple drive mirroring for redundancy. When they're set up for performance, ATA RAID adapters run RAID Level 0, which incorporates data striping. Unfortunately, RAID 0 also sacrifices reliability, such that if one drive fails, all data is lost. With RAID 0, performance scales up with the number of drives you add to the array. If you use four drives, you don't necessarily have four times the performance of a single drive, but you can get close to that for sustained transfers. Some overhead is still involved in the controller performing the striping, and issues still exist with latencythat is, how long it takes to find the databut performance with RAID is higher than what any single drive could normally achieve.

When they're set up for reliability, ATA RAID adapters generally run RAID Level 1, which is simple drive mirroring. All data written to one drive is written to the other. If one drive fails, the system can continue to work on the other drive. Unfortunately, this does not increase performance at all, and it also means you get to use only half of the available drive capacity. In other words, you must install two drives, but you get to use only one (the other is the mirror). However, in an era of high capacities and low drive prices, this is not a significant issue. If you want to eliminate a lot of bulky cable, consider SATA RAID, which uses the narrow SATA cables shown in Figure 6.1, earlier in this chapter.

Combining performance with fault tolerance requires using one of the other RAID levels, such as RAID Level 3 or RAID Level 5. For example, virtually all professional RAID controllers used in network file servers are designed to use RAID Level 5. Controllers that implement RAID Level 5 are more expensive, and at least three drives must be connected. To improve reliability, but at a lower cost, many of the ATA RAID controllers enable combinations of the RAID levelssuch as 0 and 1 combined (also known as RAID 10). This usually requires four drives, two of which are striped together in a RAID Level 0 arrangement, which is then redundantly written to a second set of two drives in a RAID Level 1 arrangement. This enables you to have approximately double the performance of a single drive, and you have a backup set, in case one of the primary sets fails. Many recent servers include four SATA host adapters with RAID functionality, enabling SATA RAID 0+1 implementations.

Today, you can get PATA or SATA RAID controllers from companies such as Arco Computer Products, Iwill, Promise Technology, and HighPoint. A typical low-cost ATA RAID controller enables up to four drives to be attached, and you can run them in RAID Level 0, 1, or 0+1 mode. Remember that performance suffers somewhat when you run two drives (master/slave) on a single PATA channel because only one drive can transfer on the cable at a time, which cuts performance in half. Four-channel PATA RAID cards are available, but most new RAID cards are moving to SATA, which doesn't have the master/slave channel sharing problems of PATA. SATA RAID cards use a separate SATA data channel (cable) for each drive, allowing maximum performance. You should use SATA RAID cards for best performance.

If you are looking for an ATA RAID controller (or a motherboard with an integrated ATA RAID controller), you should look for the following:

• RAID levels supported (most support 0, 1, and 0+1, although some ATA RAID 5 card products are now available)

• Two or four channels (for best performance, you should use four channels with four devices using PATA)

• Support for ATA/100 or ATA/133 speeds for PATA

• Support for 33MHz or 66MHz PCI slots (some servers have 66MHz slots)

• SATA RAID implementations for maximum performance, ease of installation, and reliability

If you want to experiment with RAID inexpensively, you can implement RAID without a custom controller when using certain higher-end (often server-based) operating systems. For example, the Windows NT/2000 and Windows XP or Windows Server 2003 operating systems provide a software implementation for RAID, using both striping and mirroring. In these operating systems, you use the Disk Administrator tool to set up and control the RAID functions, as well as to reconstruct the volume after a failure occurs. Normally, though, if you are building a server and want the ultimate in performance and reliability, you should look for SATA or SCSI RAID controllers that support RAID Level 3 or Level 5.

Tip

The best solutions for RAID are SATA and SATA RAID implementations that are native to the motherboard chipset's South Bridge or I/O controller hub chip. When Windows 2000 Server or Windows Server 2003 is installed on a system with native SATA or SATA RAID support, there is no need to press the F6 key when you're prompted to install a driver from a floppy disk. With add-on SATA and ATA/SATA RAID add-on cards or motherboard implementations using a separate SATA host adapter chip, you must press F6 when prompted to install the appropriate driver from a floppy disk.

The latest type of RAID array, RAIDⁿ, developed by Inostor, a division of Tandberg Data, combines special software that can specify the number of drives used for replacements. A RAIDⁿ array designed to protect against the failure of up to four drives uses only 7 drives (4 for data and 3 for parity), compared to 10 drives for a RAID 5+1 array (4 data, 1 parity; 4 mirrored data; 1 mirrored parity). Currently, RAIDⁿ arrays are available only in Inostor network attached storage (NAS) devices, but RAIDⁿ software is designed to be portable to other operating systems.

For more about RAID, see "Introduction to RAID," p. 600.

Using NAS

As an alternative to expanding internal server storage by adding ATA, SATA, or SCSI hard disks to a server (a concept often referred to as direct attached storage [DAS]), you can connect a network attached storage (NAS) device to your network. Each NAS device has its own IP address, and an NAS device contains one or more hard disks that can be accessed directly from the network.

The first company to develop an NAS device was Auspex, which introduced the first NAS shortly after being founded in 1987. Although the company went bankrupt in June 2003, its technology was sold to other firms. Using NAS devices continues to be a popular method of expanding network storage, and many companies now build NAS devices.

In addition to hard disks, an NAS device contains one or more processors to handle I/O, network, and data storage and retrieval tasks. Although a given NAS device might contain a processor similar to that used by a desktop PC or server, an NAS device is not a PC or a server. Instead, it is a specialized storage appliance.

The benefits of expanding network storage by using an NAS device include the following:

• No server downtime You can connect an NAS device to the network without disrupting existing servers.

• Easier configuration In most cases, you can configure the NAS from a web browser on a client PC. The client PC logs in to the NAS device's built-in webserver. Some NAS devices use Windows Storage Server 2003.

• Easy rack-mounting Many NAS devices fit in a 1U or 2U form factor.

• Easy backup Some NAS devices, including autoloaders, support direct disk-to-disk or disk-to-tape backups. Some NAS devices can be used as network backup appliances (NBAs) for other servers.

• Wide range of RAID options RAID options vary by product and vendor, but most vendors offer RAID 0, RAID 1, RAID 0+1, RAID 4, and RAID 5. Some offer additional RAID options.

The following are the shortcomings of NAS storage:

• The speed of storage is limited by the speed of the network For best results, an NAS should be connected to a Gigabit Ethernet network, with all clients using Gigabit Ethernet adapters.

• As the name implies, NAS is for network storage only If you use thin clients that remotely store applications, applications must be stored on a server, not on an NAS.

• NAS storage can be more difficult to secure than server-based storage For maximum protection, NAS devices should be protected by antivirus programs made especially for NAS, such as Symantec AntiVirus ScanEngine. Also, networks that use NAS devices should be protected from public access by firewall devices.

If you need speeds significantly faster than NAS can provide (about 32MBps sustained transfer rate with a Gigabit Ethernet network) and prefer more security options, you should consider a storage area network (SAN) installation.

For more about SANs, see Chapter 12, "Storage Area Networks."