What is RAID?

1. Definition

RAID is the acronym for “Redundant Array of Independent Disks”, it refers to a data replication system providing increased security, performance and failure tolerance. RAID technology originated in California, it was first imagined by a team of researchers from the University of Berkeley in 1987. It was originally a way to recognize several hard drives as a single entity. by the system. The results of the first tests show good performance but suffer from poor reliability. Efforts are then focused on the use of redundant architectures ensuring better fault tolerance. From 1988, the concept of RAID is established, the different types of RAID are defined and documented, each with its own advantages and disadvantages.

2. Difference between RAID and SLED

SLED stands for "Single Large Expensive Disk" and is based on the use of a single high performance, large capacity disk. The need to use a high performance disk firstly makes this system quite expensive. But that's not the only downside. Indeed, in the case of disk failure, the only way to reassemble the data is to restore the most recent backup. This requires a significant amount of time (add to that the time to replace the defective drive) which is not acceptable for a company's information system. Conversely, in the case of RAID, a failure of one disk does not cause the entire cluster of servers to become unavailable. RAID does not necessarily require the use of high performance disks, which represents a significant saving of money. Finally, after the replacement of defective equipment, the content is automatically rebuilt without impacting the operation of the system and therefore allows business continuity.

3. The different types of RAID

1) RAID 1

RAID 1 is akin to mirroring, which is simply the replication of data from one drive to another. Often based on just two drives, RAID 1 is relatively easy to set up.

- Easy to use
- Economical system because it does not require any particular power
- Requires little time to get the system up and running again

- 50% of the resources of each disk are used to store redundant data. This therefore represents a significant loss of volume.

2) RAID 5

Based on the distribution of data over several disks, (at least 3), RAID 5 segments the data into several parts and also includes a part called "parity". The parity is calculated from the different data packets and makes it possible to reconstruct missing data. With this system, it is easy to recalculate parity from data and vice versa to reconstruct missing data from parity and other data packets.

- Negligible loss of space compared to RAID 1
- Faster reading of data

- The system becomes fragile during the reconstruction of the RAID

3) RAID 0

RAID 0 does not secure the data itself. In the event that one of the disks falls, the entire infrastructure falls with it. The main objective of RAID 0 is to provide more storage volume at a lower cost.

- By using multiple disks, resources are multiplied by the number of disks available. Thus, with two hard drives, the reading and writing speed is multiplied by 2.

- RAID 0 does not provide real security for your data.

4. The difference between software RAID and hardware RAID

1) Hardware RAID

As part of a hardware RAID, operations management is provided by a card or a hardware component. This component called "RAID Controller" allows a total virtualization of the storage system.

The advantages of hardware RAID
- Hardware RAID controllers allow fault identification, "hot" swapping of faulty units as well as rebuilding transparently for the user
- The system load is lightened
- The system manages the various maintenance, verification and consistency operations in the background without requiring additional resources.

The disadvantages of hardware RAID
- Hardware RAID is by definition more expensive than software RAID because it requires additional components.
- The RAID controller can also fail (the firmware it embeds can also contain errors)
- Hardware RAID can cause compatibility problems between different manufacturers of RAID controllers.
- Hardware RAID offers reduced flexibility because RAID controllers are often specialized for a single type of device.

2) Software RAID

Software RAID employs much the same operation except for RAID volumes, which are managed by a software layer. Software RAID is most often used on small servers that do not always allow management by a hardware controller. It provides less security although not negligible.

The advantages of software RAID
- Software RAID is less expensive because it does not require additional components
- It offers great flexibility of administration, the only requirement being a homogeneity of the operating system

In conclusion, several types of RAID exist, they each have their advantages and disadvantages. It is therefore essential to first identify your main problem and then choose the type of RAID that will suit you best. Regardless of your choice, however, it is important to emphasize the critical importance of planning a RAID system. Indeed, it represents an investment in terms of time and money, but the benefits provided are invaluable in the event of your machine failing.


Add new comment