Lec-114: What is RAID? RAID 0, RAID 1, RAID 4, RAID 5, RAID 6, Nested RAID 10 Explained

Summary

This educational video delves into RAID technology, explaining its significance in data storage for performance and security. It covers various RAID levels, including RAID 0 for performance via data stripping, RAID 1 for data security through mirroring, and RAID 5 for a balance of both by distributing parity across disks. The script uses real-life examples like Facebook's outage to emphasize the importance of data availability and introduces RAID 6, which offers double parity for even greater fault tolerance. The video is tailored for students, professionals preparing for exams or interviews, and anyone interested in data storage solutions.

Takeaways

• 💾 RAID stands for Redundant Array of Independent Disks (or Inexpensive Disks), which is a method of storing data across multiple hard drives.
• 🔄 RAID 0, or data stripping, splits data into pieces and stores them across different disks to enhance performance but offers no data redundancy.
• 🔄 RAID 1, or mirroring, creates exact copies of data on separate disks to ensure data availability and security in case of disk failure.
• 🔄 RAID 0+1 (or 1+0) combines both data stripping and mirroring to balance performance and data security.
• 🔄 RAID 3 stores data in blocks with a dedicated parity disk, allowing recovery from a single disk failure but can be a bottleneck due to high utilization of the parity disk.
• 🔄 RAID 4 is similar to RAID 3 but distributes parity information across all disks to prevent bottlenecks.
• 🔄 RAID 5 distributes parity information across all disks, ensuring equal utilization and enhancing performance compared to RAID 3 and 4.
• 🔄 RAID 6 calculates two parities and stores them across all disks, allowing for recovery from up to two disk failures.
• 💡 The choice of RAID level depends on the balance between performance, data security, and cost, with each level offering different advantages and trade-offs.
• 🌐 Real-world examples, such as Facebook, WhatsApp, and Instagram outages, highlight the importance of data availability and the potential financial and credibility impact of downtime.

Q & A

What does RAID stand for?

RAID stands for Redundant Array of Independent Disks, or sometimes Redundant Array of Inexpensive Disks.

Why is redundancy important in RAID?

Redundancy in RAID is important for duplicating data across multiple disks to ensure data availability and fault tolerance, which helps in case one or more disks fail.

What are the two main factors companies consider when using RAID?

Companies consider performance, which includes fast read and write speeds, and security or availability, ensuring data is accessible 24/7.

What is the significance of the Facebook, WhatsApp, and Instagram outage mentioned in the script?

The outage of these platforms for 6-7 hours resulted in a significant loss of revenue and credibility, highlighting the importance of data availability and the impact of downtime on businesses.

What is RAID 0 and how does it improve performance?

RAID 0 is a level where data is broken into pieces and distributed across multiple disks. This data striping allows for parallel reading and writing, which increases performance and throughput.

How does RAID 1 provide data security?

RAID 1, also known as mirroring, involves creating exact copies of data on separate disks. This ensures that if one disk fails, the data is still accessible from the other copies.

What is the difference between RAID 1 and RAID 0+1?

RAID 1+0 (or 0+1) is a nested RAID level that combines both mirroring and striping. It mirrors data across multiple sets of disks and then stripes the data within each mirror set, offering both performance and data security.

What is RAID 3 and how does it use parity for data recovery?

RAID 3 is a level where data is divided into blocks and distributed across multiple disks, with a dedicated disk for storing parity information. Parity is used to recover data if one disk fails, but if two disks fail, data recovery is not possible.

How does RAID 4 differ from RAID 3?

RAID 4 is similar to RAID 3 in that it also uses block-level striping and parity for data recovery. However, the main difference is that RAID 4 stores the parity block across all disks, rather than on a single dedicated parity disk.

What is RAID 5 and how does it distribute parity?

RAID 5 distributes parity information across all disks in the array, rather than storing it on a single disk. This prevents any single disk from becoming a bottleneck and ensures that all disks are utilized equally.

What advantage does RAID 6 offer over RAID 5?

RAID 6 offers the advantage of having two parity blocks, which allows for the recovery of data even if two disks fail simultaneously, providing an additional layer of fault tolerance.

Outlines

00:00

💾 Introduction to RAID

The video begins with an introduction to RAID, which stands for Redundant Array of Independent Disks or Redundant Array of Inexpensive Disks. The presenter explains that RAID is used to store data across multiple disks, which are independent of each other, meaning the failure of one disk does not affect the others. The term 'inexpensive' refers to the cost-effective nature of disks compared to other forms of memory like registers or RAM. The importance of RAID is highlighted in terms of its relevance for competitive exams, college studies, and interviews. The video emphasizes the need for both performance and data availability, using the example of Facebook, WhatsApp, and Instagram's downtime causing significant financial losses. The presenter also touches on the increasing storage capacities in laptops as an indicator of the cost reduction in disk technology.

05:01

🔄 RAID 0: Data Stripping for Performance

The second paragraph delves into RAID 0, also known as data stripping. In RAID 0, data is broken down into pieces and distributed across multiple disks, which can improve read and write performance as these operations can be done in parallel. The video uses a diagram to illustrate how data is split and stored across different disks. The benefit of RAID 0 is its ability to enhance performance and throughput. However, it does not provide data redundancy, meaning if one disk fails, the entire data set is lost. The video also mentions RAID 1, which is mirroring, where data is duplicated across disks to ensure availability and security, contrasting it with RAID 0's focus on performance over redundancy.

10:03

🔄 RAID 1 and RAID 1+0: Mirroring and Nested RAID

This paragraph discusses RAID 1, which involves mirroring data across two or more disks, providing data security and availability. If one disk fails, the data can still be accessed from the mirror. The video points out that while this does increase cost due to the need for additional disks, the use of inexpensive disks makes it a viable option. The concept of RAID 1+0, or nested RAID, is introduced as a combination of RAID 0 and RAID 1. It involves both mirroring and striping data, providing both high performance and data security. The video explains that this configuration is commonly used in email and web servers due to its balance of performance and reliability.

🔄 RAID 3, 4, 5, and 6: Advanced RAID Levels

The final paragraph covers more advanced RAID levels, starting with RAID 3, which involves block-level striping and a single parity disk. The parity disk allows for the recovery of data if one disk fails, but if two disks fail, data recovery is not possible. RAID 4 is similar to RAID 3 but addresses the issue of a potential bottleneck with the parity disk by distributing the parity information across all disks. RAID 5 improves upon RAID 4 by distributing the parity information, preventing any single disk from being overused. Lastly, RAID 6 introduces double parity, allowing for the recovery of data even if two disks fail. The video concludes by emphasizing the importance of understanding these RAID levels for various applications and potential interview questions.

Keywords

💡RAID

RAID stands for Redundant Array of Independent Disks (or Inexpensive Disks). It is a data storage virtualization technology that combines multiple physical disk drive components into one logical unit for the purposes of data redundancy, performance improvement, or both. In the video, RAID is the central theme, with various RAID levels discussed to explain how they provide different levels of data security and performance.

💡Redundancy

Redundancy in the context of RAID refers to the duplication of data across multiple disks to ensure data availability in the event of a disk failure. This concept is crucial for understanding RAID's purpose, as it directly relates to the video's discussion on data availability and fault tolerance. For example, RAID 1 uses mirroring to create a redundant copy of data on another disk.

💡Performance

In the video, performance refers to the speed and efficiency of data read and write operations. RAID configurations like RAID 0 are designed to enhance performance by striping data across multiple disks, allowing for parallel operations that increase throughput. The script mentions how companies value performance for quick data access and manipulation.

💡Availability

Availability in the context of the video refers to the constant accessibility of data, which is critical for business operations. The video uses the example of Facebook, WhatsApp, and Instagram's downtime to illustrate the importance of data availability and how it impacts business continuity and revenue.

💡Data Striping

Data striping is a RAID technique where data is divided into blocks and distributed across multiple disks. This method is used in RAID 0 to enhance performance by allowing simultaneous read and write operations on different disks. The video explains how striping can lead to faster data access but does not provide data redundancy.

💡Mirroring

Mirroring, as discussed in the video, is a RAID technique where data is copied exactly onto another disk. This is used in RAID 1 to provide data redundancy, ensuring that if one disk fails, the data can still be accessed from the mirror. The video emphasizes the balance between cost and data security that mirroring provides.

💡Parity

Parity in RAID is a method of error detection and correction where a parity bit or block is calculated from the data and stored separately. RAID levels like RAID 3, 4, and 5 use parity to enable data recovery in the event of a single disk failure. The video explains how parity works and its role in maintaining data integrity.

💡Bottleneck

A bottleneck in the context of RAID refers to a point of congestion where the system's performance is limited by a particular component. The video mentions that in RAID 3, the parity disk can become a bottleneck because all read and write operations that affect parity require access to this disk, potentially slowing down the system.

💡RAID 5

RAID 5 is a RAID level that combines the striping of RAID 0 with distributed parity. The video explains how RAID 5 distributes parity across all disks, preventing any single disk from becoming a bottleneck and improving overall performance and fault tolerance compared to RAID 3 and RAID 4.

💡RAID 6

RAID 6 is a RAID level that extends RAID 5 by calculating two parity blocks. This provides an additional layer of fault tolerance, allowing the system to sustain the failure of two disks without data loss. The video highlights RAID 6's advantage in high-availability storage systems where data integrity is critical.

Highlights

Introduction to RAID (Redundant Array of Independent Disks) and its importance for data storage.

Explanation of the term 'redundant' in RAID and its significance for data duplication.

Discussion on the cost-effectiveness of disks and their role in RAID configurations.

Importance of RAID for performance and data availability in organizations.

Real-world example of Facebook, WhatsApp, and Instagram outage highlighting the need for RAID.

Description of RAID 0 and its method of data stripping for improved performance.

Advantages of RAID 0 in terms of read and write speed due to parallel operations.

Introduction to RAID 1 and its mirroring technique for data redundancy.

Cost-benefit analysis of using inexpensive disks for RAID 1 configurations.

Explanation of data security and availability in RAID 1 with real-world examples.

Combination of RAID 0 and RAID 1 to form RAID 10, offering both performance and data security.

Description of RAID 3 and its block-level data stripping with parity for fault tolerance.

Challenges with RAID 3, such as the single parity disk becoming a bottleneck.

Introduction to RAID 4, which is similar to RAID 3 but with distributed parity.

Explanation of RAID 5, which distributes parity across all disks to prevent bottlenecks.

Description of RAID 6, which calculates two parities to allow recovery from the failure of two disks.

Practical applications of different RAID levels in email and web servers.

Summary of RAID levels and their respective advantages for data storage solutions.

Transcripts

Dear student, welcome to Gate Smashers

In this video I'm going to explain

RAID

That is redundant array of independent Disk

or redundant array of inexpensive disk

So guys here I'll discuss about different RAID levels

with real life examples

So this video from competitive exams point of view, for college and university and for interviews

is very important

So guys so like the video and subscribe the channel

if you havn't done and if you have done so do subscribe from other other device

Subscriber are very important

lets start first of all

here we are writing Redundant array

redundant means

redundant means copy

it means duplicacy

means multiple copies of 1 thing

so here

reduntdancy

duplicacy

that is of

that is of disk

and why we're using disk

we're using disk for storing the data

so we are storing data in disk

that disk is independent

lets say if I've 5 disk

and array mean

that we don't have 1-2 disk but we've multiple disk

so this disk

first this is independent

it means it is not like that if 1 is closed

then rest will also get automatically closed, there is no dependency between them

and second why it is called inexpensive

because if we talk about memory hierchary

it is said that register are very much expensive

then cash

then RAM

So somewhere this memory are more costly

but we talk about disk

So disk at this time, its cost is less

and even gradually it is getting decrease

when you've purchased your laptop

so you also have seen that first there is 256 and then there 512 GB

then 1 TB and now a days if you buy a normal laptop theres is by default 1TB is coming

it may be like you can take 2 TB in future also

So it mean that we're using indepentent and inexpensive disk we're using

for storing the data

but why we're doing

what is need of this, this is the first question that arise

So guys whenever we think of company or organsition point of view

so we need 2 things

first is performance

performance in case of read and write

means you from data

you want to read data, want to read data from disk

and write

means want to change data in disk

some transaction has changed something

you want Read and Write

to be fast

so you want performance

Second you want security

avialability

that data should be avaialalbe 24*7

and its real life example I'll give to you

you have heard fews days ago

that facebook, whatapp and insta

was down for 6-7 hours

and you might have heard this also

and because of that Mark zuckerberg income

how much loss does he faced

he faced 7 billion loss

7 billion dollar loss

from 3rd place he has shifted to 5th place

only becasue of 6-7 hours

services were down

and what is there in that services

what we do in whatsapp

sharing data only

what is there in facebook

our profile and friends profile

that is also data

and if we talk about instagram

in that also there is only photos and videos, that is also data only

so it is dealing with data only

and only because of this small down

so because of that there is 7 billion of loss

so guys

thats why companies want

there data should be 24*7 available

and if is not available it would be great loss for company

so that loss

can be in term of money also

and it can be in term of cerdibility also

means trust of them

I'll give one more example of this

if you've heard about telegram

So telegram in only those 6-7 hours

60-70 millions accounts has been added

when facebook and all were down

that 6-7 hours

60-70 new accounts has been created

So see one compant faced a good benefit and another one has faced loss

and only because of avaialbity and performance

so lets start, so you understood storty

that why we are using and what is its purpose

So we'll talk anout different levels

So in Raid we've diffrenet levels

that how we are keeping the data

So first level in this is RAID 0

What we do in it

data stripping

means we are breaking data, breaking data in pieces

and keep in different disk

like you can see in diagram

that in RAID 0, original data that I've lets say it is A

B,C,D

means you can put data in this way

what I did, I broke A in A1 and A2

B into B1 and B2 means you can do mutiple pieces also

but if you see in simple diagram

So I break A1 into A1 and A2

then on A3 and A4 and A5 and A6

I mean to say that breaking data in pieces

and keeping that data in different disk

so what is the use of this

its benefit is performance

RAID 0 gives us

performance

because if you want to read, you want to read data of A

so parallelly you can read the data from both as it is independent disk

parallelly you can read data from this

So your performance will be fat

same if you want write(B)

you want change something in A7

and in A8

so both write will be parallelly

so this give high performance and throughput

RAID 0

So remember this point

then

so it can be asked from you that next level is RAID 1

What is purpose of RAID 1, it is called mirroring

as RAID 0 is called data stripping

here it is called mirroring

here we are not breaking data

same data copy

is shifting to another disk

means I've a data A,B,C

I've made a mirror of that, mirror means like there is mirror image

copy of that data

is kept on another disk also

So you will ask sir it will cost a lot

cost is there

but again inexpensive disk

means we're choosing that disk or memory

whose price is not that much high

we are not using ssd, cash, RAM

we 're using that disk

whose performance is also good

and price is also not that much high

So we're keeping multiple copies

so you got its simple advantage

it is not pointing out on performance

here it talks about availability

data security

lets say if one disk failed

still I can access that data because in other disk that work is going

this is 2 simple example

companies keep 3-4 mirrors and distribute in whole globe

if there is problem in one geographical area

So it can fetch data from different geographical area

but idea must have clear from here

So mirroring work is to secure data in a way

in case of

failure we can access data from another

and in case if this is coming in your mind that if both disk fails

So guys there is no solution of this, you can give 10 mirrors

then also you can say and what if 10 mirror fails

so that is a worst case

but still we are reducing the probability

if failure occurred

whether I'll able to access my data or not

so we are increasing its probability

that yes I can still access my data

so we're improving that thing

so this is combination

means here you can see RAID 1+0

at same places 1+0 can also be written

so we don't call it 10

it is not 10 but is 1+0

1+0 means we've mix 1 and 0

So 0+1 can also be written on 1+0 also

but what we did in 1+0

that data is here mirrored also

and along with this it is striped also

see A1 and A2

means we break the data also, original data was A

it is broken in 2 parts A1 and A2

and kept a copy of that also

and its further part also made A3 and A4

and kept a copy of that, means stripping is also done like data is broken

and mirroring also

that’s why we say

nested raid because both combination

and it is advantageous

very useful

and real life application

whether you talk about email server or web server

there it is mostly used

because this raid 0 it give high performance

so it is using that also

and Raid 1 data security

so it is using this also

so it ls the combination of both

then next if we talk about RAID 3

RAID 2 is also there but has become absolute in that we break data in bit level

that is not used here

Now we'll come to RAID 3

in this we break data in block level

I've data A

I've divided that in bloc

blocks means

broken data in pieces

in pieces

and capture data in different disk

A1 here and A2 here and A3 here

A4 here A5 here and A6 here

Breaking original data into 6 pieces and kept it like this

So in last you might be seeing that what it is A

this is parity bit, means taking parity of data

and stored in a different disk

all data

what we need to dot that along with storing original data

its parity is also stored, why we do this

parity storing means

if one of the disk is failed

if one of the disk is failed

Still I can recover the data from

parity

I'll give ls simple example

lets say data is 1

and this is 2

and this is data 3

just for an example, I'll tell you with a simple example

so if anyone ask you can easily answer that

So method of finding parity is adding all of three

So 1+2+3 will be

it will be 6

so parity 6 we've stored here

so like this we have find out its Parity

now lets say in future A1 fails

A1 disk fails

so now how I'll recover the data

with the help of parity

Parity is 6

and I know how parity was found 1+2+3

so what I'll do

I've to find this one

take rest of the parity

its sum is 5

So will subtract from parity so it will be 5

So see its data is 1

this I'm telling in a form of example

that if 1 disk is failed

Still I can recover the data from the parity

only 1 disk but if 2 disk fail

then we can't do anything

then we can't do anything

then we can't find with parity as there will be confusion

then multiple options can come, so here

So here you got the advantage that we broke block level data

in blocks

made blocks of original and made one parity

but what is the problem here

problem is that all parity is there in 1 parity

data is in hard disk

kept parity in different disk

problem from this is that in case

in case if there if problem came in parity

So all the parities will not able to get use

second bottle neck can come

if anyone read

or write, lets say here also write and here also

and here also

whenever you write or change

so you need to change in parity also

Parity will change always

lets say if I've here instead of 2 I've done 20

then new parity I've to calculate

so whenever you'll write

then you've to take parity from here

So this disk will be used a lot

because of this it can came in bottle neck state

it will be very usable that it can create problem in accessibility

So this problem can come and this problem is solved by next one

RAID 4 says

what we're doing in RAID 4

same concept

but in this we've saved parity in such way

in actual there was not much difference in RAID 3 and 4

but this problem is solved by

RAID 5

what RAID 5 did

All these parities

are distributed

Parities are distributed

Parity in not in 1 disk

Now that parities are distributed all disk

now there is data and parity both

so what will happen because of this

because of this no disk will be over utilized

all disk will be equally utilized, right operation will be performed well by this

as compare to RAID 3 and 4

finally there is RAID 6, this is last,in raid 6 there is little difference

in this we are calculating 2 parities

instead of 1 we are calculating 2 parities

A and A

this data A1,A2 and A3

calculated 2 parities of this

calculating 2 parties means as in earlier one if 1 disk fails

I can recover from 1 parity

but if 2 disk fail

then it is not possible

but here advantage is

that if 2 disk also fails

still I can recover the data from the parity

because I've 2 parities

So there are 2 equations

so from 2 equations I can find out the value, like we normally do

that in one equation x, y value is this and in second it is this

so obviously we

find out them

I mean to say

That in RAID six advantage is

that from 2 parities if 2 disk fails

then that you can easily recover

So from all these parameter whatever question comes you can easily explain

Thank you