Fault Tolerance

Running Head: FAULT TOLERANCE: RAID TECHNOLOGY

Fault Tolerance: RAID Technology

Introduction

We rely on computers everyday, not only for business, but in our personal lives as well. It is virtually impossible to go an entire day without interacting with some form of computer system. From driving a car to shopping for groceries, the presence of computers is everywhere. Society has evolved to the point where some of our daily processes would no longer be acceptable if it had to be done manually. A prime example is the US postal service. The Internet, email, and instant messenger services have virtually replaced the traditional method of sending and receiving mail. We have not only become accustomed to the features and speed of computers, but we depend on their reliability and availability in order to accomplish our day-to-day tasks.

When these systems are unavailable due to technical difficulties, the loss is immediately noticed. The loss of system functionality can even be life threatening when that system is assisting someone to breathe. In business, the loss of a database server due to a failed hard drive can cost millions of dollars in data loss. This is more than a mere inconvenience; downtime can cost a life Ð'- the life of a human, and the life of a business.

What is Fault Tolerance?

Fault Tolerance is a characteristic of a system that will allow it to continue functioning in the event that part of the system fails. This usually involves having multiple critical components running parallel. These components can include CPUs, power supplies, cooling fans, memory, or hard drives.

What is RAID?

RAID stands for Redundant Array of Inexpensive (Independent) Disks. In a RAID implementation, multiple disks are used to simulate one large logical drive. Commonly used in servers and high performance computers, RAID is used to help protect data, increase disk storage performance and depending on the level of RAID used, provide fault tolerance. RAID works by consistently spreading data across each drive in the array. The data is then broken down into chunks and written to the hard drive based on the current level of RAID in use.

RAID requires the use of software or hardware to function properly. From an end user point of view, there is no difference between hardware and software RAID. Hardware RAID includes the use of dedicated hardware to control the disk array. It is independent of the operating system in use and does not consume system resources. A hardware RAID controller or device uses its own processor and memory in order to function, leaving resources available for other running applications. Software RAID is limited and dependent on the operating system. Just like any other software application, software RAID consumes system resources (memory and CPU cycles) that are used to run other applications. This eventually decreases the overall performance of the server.

RAID technologies offer various benefits depending on the level used. All levels of RAID offer one or more of the following features; increased performance, reliability, data protection, and fault tolerance. Below are some of the more popular levels of RAID along with some of their individual advantages and disadvantages.

Levels of RAID

RAID 0 - Raid 0 reads and writes data on all disks simultaneously

Requires: Minimum of 2 drives

Advantages: High transfer rate

Improves speed

Full capacity is available

Disadvantages: No redundancy

Less reliable, if one disk fails, you lose the whole array.

RAID 1 - Raid 1 makes an identical copy of one disk on another

Requires: Minimum of 2 drives

Advantages: Easy to implement

Easy recovery from failures

Achieves redundancy through duplication

Very reliable

Disadvantages: No increase in performance

Storage capacity is as large as the smallest drive

Slower than RAID 0 at twice the cost

Half the capacity of RAID 0

RAID 3 - Provides byte-level striping with a dedicated parity disk

Requires: Minimum of 3 drives

Advantages: High data transfer

Disadvantages: Single parity drive does not support multiple, simultaneous read and

write requests.

RAID 4 - Similar to RAID 3, but uses block level striping.

Requires: Minimum of 3 drives

Advantages: Data striping supports multiple simultaneous read requests.

Disadvantages: The parity disk can create write bottlenecks.

RAID 5 - Data are striped across three or more drives for performance, and parity bits are used for fault tolerance.

Requires: 3 or more disks

Advantages: High performance

High data protection

Supports multiple simultaneous reads and writes

Fault tolerance