RAID CONCEPT

RAID Concepts

Engenio Storage GroupInternal RAID

2008

LSI Confidential 2

RAID Concepts

• Objective: At the end of this presentation you will be able to …

– Discuss basic RAID technology market positioning

– Explain the inherent advantages of using RAID

– Identify appropriate RAID levels to use for specific platforms and applications

LSI Confidential 3

What is RAID?

• RAID stands for an redundant array of independent drives• Intelligently manages drives in unison performing data read/write

algorithms across drives delivering– High levels of protection against downtime and data loss (mirroring)– Larger storage volumes achievable– Improved performance (striping)

LSI Confidential 4

RAID Level Overview

• Selecting the proper RAID level for a specific data storage application requires consideration be given to the benefits of each

LSI Supported• RAID 0 (1 to 32 disks)• RAID 1 (2 disks)• RAID 5 (3 to 32 disks)• RAID 10 (4 to 16 disks)• RAID 50 (6 to 60 disks)• RAID 6

Use• Video/Audio streaming• OS boot• Transaction/Web server• Database• Data warehousing• Large capacity disk arrays

LSI Confidential 5

RAID 0 - Striping

Definition

Striping - writes data across multiple drives Involves partitioning each drive storage space into stripes that can

vary in size from 2 KB to 1 MB.

These stripes are interleaved in a repeated sequential manner.

BenefitsHigh data throughput, especially for large filesNo capacity loss penalty for parity.

Drawbacks No fault tolerance: If any drive in the array fails, all data is

lost

Uses Non-critical data requiring high performance Environments that do not require fault tolerance

Drives Minimum 1 / Maximum 64

Fault Tolerance No

LSI Confidential 6

RAID 0

LSI Confidential 7

RAID 1: Mirroring

Definition

Mirroring - writes duplicate data to more than one (usually two) drives

Protects against data loss in the event of a device failure

Benefits Provides 100% data redundancy Should one drive fail, the controller switches reads and

writes to the other drive.

Drawbacks Expensive: Requires two drives for the storage space of

one drive Reduced Performance during drive rebuilds.

Uses If availability is critical

Use RAID 1 for small databases or any other environment that requires fault tolerance but small capacity

Drives 2

Fault Tolerance Yes

LSI Confidential 8

RAID 1

LSI Confidential 9

RAID 5 : Striping with Rotational Parity

Definition

Striping with rotational parity - blocks of data and parity information is stripped across all drives

RAID level 5 is the most popular configuration, providing striping as well as parity for error recovery

Benefits Uses one disk worth of space to achieve data redundancy

If a hot spare is available, it can perform a rebuild automatically

Drawbacks

Cannot match RAID 0 in write performance due to processing required to compute parity

While a controller is rebuilding a drive, users will experience reduced performance if reading or writing data

Uses

Any application that has high read request rates and average write request rates

Transaction servers, web servers, data mining applications, exchange servers

Drives Minimum 3

Fault Tolerance Yes

LSI Confidential 10

RAID 5

LSI Confidential 11

RAID 10 : Spanning two RAID 1’s

DefinitionSpanning two RAID 1’s - writing duplicate data to more than one pair of drives to protect against data loss in the event of a up to two disk failures (one per array)

Benefits Optimized for both fault tolerance and performance Provides both high data transfer rates and complete data

redundancy

Drawbacks Requires half the available disk space for data redundancy

Same as RAID level 1.

Uses

Environments that require 100% redundancy of mirroring (RAID 1) and the enhanced I/O performance of stripping (RAID 0)

Ideal for smaller organizations needing a high degree of fault tolerance and moderate to medium capacity.

Drives Minimum 4 / Maximum 16

Fault Tolerance Yes

LSI Confidential 12

RAID 50 : Spanning two RAID 5’s

Definition

Data is “striped” across multiple drive groups (super drive group). For data redundancy, drives are encoded with rotated XOR redundancy. RAID 50 provides the features of both RAID 0 and RAID 5. RAID 50 includes both parity and disk striping across multiple drives.

Benefits RAID 50 provides high data throughput, data redundancy, and very good performance.

Drawbacks Requires at least twice as many parity drives as a single RAID 5.

UsesRAID 50 works best when used with data that requires high reliability, high request rates, and high data transfer and medium to large capacity.

Drives Minimum 6

Fault Tolerance Yes

LSI Confidential 13

RAID 50

LSI Confidential 14

RAID 0+1 Enhanced Mirroring

Definition

The controller combines the performance of data striping (RAID 0) and the fault tolerance of disk mirroring (RAID 1). Data is striped across multiple drives and duplicated on another set of drives.

BenefitsOptimizes for both fault tolerance and performance. Provides excellent performance for all data needs. May be simultaneously used with other RAID levels in an array.

Drawbacks Requires half the available disk space for data redundancy, the same as RAID level 1.

Uses If a drive fails, the controller uses the parity drive to recreate all missing information.

Drives Minimum 4

Fault Tolerance Yes

LSI Confidential 15

RAID 0+1

LSI Confidential 16

RAID 1E Enhanced Mirroring

Definition

Enhanced mirroring - combines mirroring with data striping The first set of stripes are the data, and the second set of

stripes are mirrors of the first data stripe contained within the next logical drive.

Benefits

Shares the characteristics of RAID 1, but allows more than two drives, including odd numbers of drives

If one of the drives fails, the controller switches read and write requests to the remaining functional drives in the RAID level-1E array.

Drawbacks 50% storage efficiency

Uses When array availability is most important For small databases or any other environment that requires

fault tolerance but small capacity

Drives Minimum 3

Fault Tolerance Yes

LSI Confidential 17

RAID 1E

LSI Confidential 18

JBOD: Single Drive Control

BenefitsSingle drive control - the ability to combine odd size drives using all of the capacity of the drives.

Drawbacks Decreases performance because of the difficulty in using

drives concurrently or to optimize drives for different uses

Uses Works best when used if you have odd sized drives and

you want to combine them to make one big drive

Fault Tolerance No

LSI Confidential 19

RAID 6 : Striping with Dual Rotational Parity

DefinitionDistributed parity – disk striping and two independent parity

blocks per stripe Can survive the loss of two disks without losing data

Benefits Data redundancy, high read rates, and good performance

Drawbacks

Requires two sets of parity data for each write operation, resulting in significant decrease in write performance

Additional costs because of the extra capacity required by using two parity blocks per stripe

Uses

Any application that has high read request rates and average write request rates

Transaction servers, web servers, data mining applications, exchange servers

Drives Minimum 3

Fault Tolerance Yes

LSI Confidential 20

RAID 6

LSI Confidential 21

RAID Level Summary

• RAID 0: Fastest and most efficient level but offers no fault tolerance• RAID 1: Performance-critical, fault tolerant environments, but requires

2X storage• RAID 5: Best choice for multi-user environments which are not write

performance sensitive• RAID 10: Ideal for environments that require 100% redundancy with

enhanced I/O performance of stripping and can afford such an investment

• RAID 50: Works best when used with data that requires high reliability, high request rates, and high data transfer rates

• RAID 0+1: Optimal for applications needing both fault tolerance and performance. Provides excellent but additional capacity investment

• RAID 1E: Great choice for small databases or any other environment that need fault tolerance but have small capacity requirements

• RAID 6: Ideal for organizations of all sizes requiring data redundancy, high read rates, and good performance