NWA-PCUG Newsletter Article, June 2003
President's Column
by Dale Oliver, President
d_oliver@swbell.net
(click to email author)


We addressed the first topic in an upcoming series of discussions on data security at the May meeting. We introduced the concept of RAID, which stands for Redundant Array of Independent Disks. The basic strategy behind RAID is to utilize more than one disk drive to provide advantages of data security, access speed, or both. We talked in some detail about five well-known types of RAID, which are summarized in the following article.

RAID 0 is commonly known as “striping” because data is written to two or more disks simultaneously in what you can visualize as “stripes”or blocks of data. Since a single disk does not have to perform all the read/write operations, performance is substantially improved. This is a simple, easy-to-implement approach to greatly enhance drive access performance, but it offers no data protection.

    ADVANTAGES: Very fast disk read/write performance. Simple design. Easy to implement.

    DISADVANTAGES: No data protection because there is no redundant data. High risk: If only one drive fails, there is no way to rebuild the data it contained and the data on the entire disk array is lost.

RAID 1 is also known as “mirroring” and is possibly the best-known RAID configuration. Mirroring requires two or more hard drives, and relies on a hard drive controller that automatically keeps an exact copy of the hard drive’s contents. The hard drive controller is capable of writing data to multiple disks concurrently, so there is basically no performance penalty. RAID 1 is an excellent solution for home users, home office users, and small business users.

    ADVANTAGES: High data safety due to 100% redundancy of data. Data is backed up automatically in real time. Modest cost.

    DISADVANTAGES: Inefficient use of disk capacity, since an exact copy of the data is written to two or more separate drives.

RAID 5 is a configuration that requires three or more drives, and relies on parity information written to disks. This configuration allows rebuilding of data in the event of a drive failure. However, rebuilding the data is difficult in comparison to RAID 1. Also, since the parity information must be created on the fly, RAID 5 requires a complex controller setup. RAID 5 has reasonably good read/write rates, and is geared more toward business applications than personal computing applications.

    ADVANTAGES: Efficient use of disk capacity. Good read/write performance.

    DISADVANTAGES: Requires at least three disk drives. Reconstructing data after a drive failure is more difficult than RAID 1.

RAID 10 is a combination of RAID 1 “mirroring” and RAID 0 “striping.” While this configuration requires at least four disk drives, it offers the best performance, protection and capacity of all RAID configurations discussed here. In many cases, this configuration can even sustain multiple simultaneous disk failures. RAID 10 configurations are typically used in business environments that require high performance combined with high data protection, such as database servers.

    ADVANTAGES: Best combined performance, protection and capacity of all RAID configurations discussed here.

    DISADVANTAGES: Very expensive. Limited scalability.

RAID 50, as you might guess, is a combination of RAID 5 parity and RAID 0 striping. This configuration requires at least three disk drives per subset of disks, and is often configured with multiple disk drive sets. It provides a relatively high degree of fault tolerance, since one disk per subset may fail without data loss. Data rebuild speed is also fairly good since parity bits are distributed across RAID 5 hard disk sets. RAID 50 configurations are typically used in business settings when fault tolerance and disk read performance is desired.

    ADVANTAGES: Relatively high fault tolerance. Good read performance.

    DISADVANTAGES: Disk failure impacts data throughput. Data reconstruction takes longer than configurations utilizing a mirrored solution.

In summary, most of our club members are interested in home and home office computer use. With this in mind, RAID 1 is probably the best choice for most of us desiring more data safety.

Grant Wales mentioned a software product he likes called Second Copy (more info at http://www.centered.com) which allows you to create automatic backups of hard drive contents easily, but since the approach utilizes a software program running on your computer as opposed to hardware-based controllers, a performance loss will certainly be noticeable in comparison with hardware-based RAID configurations.

Also, you need to be aware that most RAID configurations are designed to protect against hard drive failure, but they do not take the place of other backup strategies that can protect you against disasters such as fires and floods, since the hard drives are located within the same computer. For protection against disasters that can completely destroy the computer, another strategy such as backing up to magnetic tape and taking the media off-site is your safest bet.

We will have a general purpose Q & A session this month, and then resume our data security topic thread next month when we discuss the basics of backing up your data.

Click here to return to top



==================================================================