RAID SYSTEM

Kanik Kehar
E-mail- [email protected]

I.M.S Engineering college Ghaziabad

Branch-IT year 3rd Roll no. 0814313025

Abstract

RAID is all about, to connect multiple disk 1 Introduction to RAID

drives and make themappear as if there
was only one disk. Then better RAID stands for Redundant Array
performance, storage capability and of Independent Disks. The idea of RAID
reliability are achieved, than can be done technology was to use an array of hard
by using just a single disk drive. disks for either better performance or
In this paper we have conducted research better security against disk failure. Raid
about RAID SYSTEMS i.e why we use can use 2 or more disks at once to increase
RAID? What is the working and how it data reading and writing speed, It can use
can be implemented with its various uses 2 or more disks to store the same data so
and criteria. disk failure will not mean that you lose
Now days there are mainly three ways of your data, or RAID can be a mixture of
accomplish a RAID system. It can be both. A RAID Array of disks will appear
either hardware driven, software driven or to an operating system as a single disk as
a combination of both. The performance of
extra storage space is not provided by
hard ware driven RAID systems are
RAID.
generally higher than software driven ones.
10One issue with RAID systems is that all
disks must have the same storing capacity 2 What RAID Does?
to work properly. If they have not, the
storage capacity of all disks will be The idea behind RAID is to take the
restricted to have the same capacity as the conventional disks in personal computers
smallest disk drive. If for we for instance and gang them together in parallel. The
have 3 disks with a capacity of 20 resulting assembly gives you the low cost
GB respectively, and one disk with a of disks manufactured in high volume plus
storing capability of 10 GB, the total good reliability and a multiplier on the
storing capacity will be 40 GB (4x10 GB), performance of individual disks.The host
since all disks will be restricted to a adapter (frequently called a controller in
maximum storage of 10 GB RAID systems) sits between one high-rate
data stream (on the computer side) and The main idea behind RAID is, as
several lower-rate streams (on the disk mentioned in the introduction, to take
side). When the computer writes to the some inexpensive disks and group them
disk, the host adapter takes high-rate data together, which will make the system see
and breaks it into multiple synchronized them as one single disk. This is done by
streams, one for each disk, in a process using a RAID controller card that handle
called striping. Reads by the computer all I/O to the disks, and which knows
cause the host adapter to take a data stream where the stored data can be found.
from each disk, multiplex the set of RAID works in three different ways to
streams into one stream, and send that provide the quality attributes mentioned
resulting stream on to the computer.In the above. These ways are
example shown below, the one high-speed 1 Mirroring
stream splits into four separate disk data 2 Striping
3 parity
streams at one-fourth the rate of the
of which each can be used either
combined stream.There are six different
separately or mixed with one or more of
levels of RAID functionality. The simplest
the others. This is why RAID is divided
RAID system, RAID level 0, merely
into different levels.
stripes the data onto multiple disks for
better performance. There is no overhead 4.1 Mirroring
for redundant data storage and no
protection against failure. The highest The easiest way to get both availability
level is RAID 5, which provides both and fault tolerance is to make a copy of all
striping for performance and redundancy data on a second disk. This is called
for failure protection. mirroring and you normally get one MB
for every two MB of physical disk space.
3 Why use RAID ? You will always have the second disk to
read from if the other disk fails. The
RAID has for a long time been something disadvantages of this method are waste of
that you only find in large server systems, disk space and that you will not get higher
but latelycheaper RAID controller card write performance. You can however get
have made it possible to get a RAID higher read performance because reads
system even for small servers and home can occur simultaneously on every drive.
computers. These will of course not have
all the features, which the more
expensive ones have. Different levels of
RAID have different advantages and
disadvantages. Therefore one must
make an analysis of the workload before
deciding what to buy. The choice also
much depend on the quality attributes
needed. Some examples of quality
attributes one can get by using a
RAID system is data redundancy, fault
tolerance, increased capacity and increased
performance11.

4 How does RAID work?

 Figure 2 showing striping technology used

4.3 Parity

Figure 1 showing the mirroring technique used Mirroring and striping are fairly easy to
understand. Parity however is a bit more
4.2 Striping complicated. In the same way as with
mirroring, it is used to improve the
Whereas mirroring and parity deal with availability but without the waste of space.
improvement of reliability, striping is used If you have X number of data elements,
to get higher performance. The idea is to they can be used to create a parity. Then
split data into small pieces, which then are you end up with X+1 data elements. It is
distributed across the disks. This way the always possible to recover a lost element
disks can work in parallel with different by using the others. The advantage with
pieces of data. You will not lose any disk parity is of course that you have no single
space as with mirroring and striping. One point of failure. However, to achieve this,
big disadvantage with this method is that if it will cost a lot of computing power.
one disk breaks, all data will be lost.
Therefore it is most often not used
alone but in combination with mirroring or
parity. Only such data that can be recreated
by the application, such as cache or other
temporary stored data is recommended to
store using only striping.

Figure 3 showing the parity technique

5 Different levels of RAID

Now when we know what RAID can do,

we will look at how it is implemented in
the different levels of RAID. The ones minimum number of hard drives required,
described below are the standard RAID is the reason this level is not used
levels. Some companies though, have today .
developed their own levels of RAID.
5.4 RAID 3
5.1 RAID 0
This level uses data striping with parity.
Raid 0 is the simplest RAID level. It uses The parity is stored on dedicated disks.
striping only to get higher performance Parity makes it possible to check the
.Therefore it is not really a RAID system, redundancy, for example, store data on two
because it has no data redundancy. disks with striping and have a third disk to
Because of that, it is not recommended to store the parity. The performance is good
store important data with RAID 0. If one but when data is written, the parity data
disk is lost, so will all data stored will slow down
on that disk be as well. the writing process.

5.2 RAID 1
5.5 RAID 4
RAID 1 is the easiest way to get data
redundancy. It uses mirroring to store two RAID 4 is very similar to RAID 3. The
copies of all data on two separate disks. difference is that it use larger blocks of
Many computer systems need high data than RAID 3 does. This makes it
availability without the requirement of possible to change the data stripe size to
more performance. In that case RAID 1 suit the applications needs. In the same
can be a good solution. Neither RAID way as with RAID 3, however, the extra
0 nor RAID 1 uses much computing parity disk will have a negative
power. The RAID controllers that handle impact on performance.
these two RAID levels are therefore
cheaper than controllers for higher RAID
levels. 5.6 RAID 5

5.3 RAID 2 This level is similar to RAID 4 in such a

way that it uses data striping with larger
RAID 2 use striping and a special kind of blocks of data. The difference is that it
redundancy technique, which is not tries to remove the bottleneck that RAID 4
described above. The technique used is bit has. That is done by skipping the dedicated
level striping with Hamming code ECC. parity disks. The parity is instead
Separate disks are used for data storage distributed amongst all disks. When
and ECC. The Hamming codes are the bottleneck is removed still one problem
calculated and written to the ECC remains. It still needs to calculate the
disks at the same time as data is written to parity, and that is still more costly than for
its specific disk. The code is calculated example mirroring. For fault tolerance and
again when data is read from the disks. performance reasons, the data and parity
This is done to check that it has not been are never stored on the same disk. This
changed since the time it means that if one disk goes down, that data
was written. The complicated and can always recovered by using data on
expensive RAID controller hardware other disks to calculate what data
needed for this level of RAID, and the disappeared. This level is one of the most
used levels today, because many think it is
the best combination of the quality
attributes such as performance and fault
tolerance.

6 Combinations of different levels

As explained above the different levels

have different advantages. Then why not
combine two levels and get the advantages
from both? Well, that was the idea behind
combinations such as
0+1, 1+0, 0+3, 3+0, 0+5, 5+0, 1+5, and
5+1. The most common one is 1+0 Figure 5 showing raid 0+1 combination
because it gives you both performance and
good data redundancy without the need of
complicated and expensive hardware that 7 RAID requirements
other combinations needs. The extra waste
of space may be cheaper than the more In order to have a RAID system on your PC you
expensive RAID controller. will need two things: a RAID controller and at
least two identical hard disk drives. If you
6.1 RAID 0+1 Striping and want to setup a system different from RAID0
Mirroring or RAID1 more hard disk drives may be
necessary, as we explained in the previous
RAID 0+1 or Striping + Mirroring as you page. On this tutorial we are assuming that
would imagine is a combination of the you are going to build a RAID0 or a RAID 1
above two setups. This setup takes the system, so we will assume a system with two
advantages of both the stripping setup and hard disk drives from now on.Nowadays
the mirroring setup. You get the increased several motherboards come already with an
performance of splitting the data across
embedded RAID controller, making it very
multiple drives, however each of these
striped drives will have a mirror as well for easy to setup a RAID system: all you will need
the data backup and security against is two identical hard disk drives (if your
failure. The obvious drawback here is the motherboard has RAID capability, of
cost involved. The minimum amount of course).So the first thing you need to check is
hard disk drives used in this configuration whether your motherboard has an embedded
is 4. This puts most home users out of the RAID controller or not. This can be seen on
equation as not only do you need to buy 4 the manual of your motherboard. The
hard disks but the PSU has to cope as well.
motherboard chipset – the south bridge chip
(which is also known as ICH, I/O Controller
Hub, on Intel chipsets) to be more exact – is in
charge of controlling the hard disk ports of
your motherboard. So the south bridge chip
of your motherboard needs to have an
embedded RAID controller. On Intel chipsets,
this chip needs to have the letter ”R“ in order
to have this feature. For instance, ICH7 chip
does not have RAID feature, while ICH7R
does. The same thing may happen with In this example we would need to use 4
chipsets from other suppliers. For example, 80Gb drives. RAID 0+1 is a combination
VIA VT8237R has RAID function while VT8237 of the two above and so storage works
doesn’t. out as a combination of the two as well.
The logical drive will appear as a single
drive, this drives capacity will be 160Gb.
The 2 striped drives will be included in
Some manufacturers call RAID function with a the logical drives space, but as above the
fancier name, like ”Intel Matrix Storage“ or mirrored drives will appear invisible to
”NVIDIA MediaShield Storage“. At the end is the user.
all the same thing.

8 Disk Capacities using RAID 9 Are there any alternatives to

RAID?
These RAID functions give you varying
capacities for your hard disk. To Now when we know what RAID is and
illustrate this we will take an example of how it works, one may wonder if there are
a user using only 80Gb hard disks. We any other techniques that do the same or
will take each of RAID levels mentioned similar thing. As mentioned earlier in this
to see what drive capacities you would document there are obvious advantages to
get out of them. RAID compared to other techniques, for
instance SLED (Single Large
8.1 RAID 0 Striping Expensive Disk). However, RAID is not
all good though. Since a RAID system is
placed at one single geographical location,
With RAID 0 and using the 2 80Gb hard it is very vulnerable to disasters. If the
disks you would get the full 160Gb of room is flooded, all data on the disks may
storage space. Although the data is split be lost. Moreover, if there is a temporary
between the 2 hard disks. There is no failure, because of a power outage, a
data redundancy (duplicate data). This hardware or software failure, etc., then the
allows for the full storage space to be data on a RAID will be unavailable for
used. the duration of the outage14. 15At the
Berkeley University of California they
8.2 RAID 1 Mirroring perform researches about alternative
solutions. Such a solution is RADD, or
Redundant Array of Distributed Disks.
When using two 80Gb hard disks with RADD: s can suppor redundant copies of
the RAID 1 function you would only data across a computer network at the
receive 80Gb of storage space. Because same space cost as RAID: s do for
you are using the two drives to contain local data. Such copies increase
the same data, the logical drive will availability in the presence of both
appear as a single 80Gb drive. temporary and permanent
failures (disasters) of single site computer
8.3 RAID 0+1 Striping systems as well as disk failures. As such,
RADD: s
should be considered as a possible
alternative to traditional multiple copy
techniques. Furthermore, RADD: s are
also candidate alternatives to high 11 Refrences
availability schemes such as hot
standbys. 16This is a new technology, and 4RAID, What is a RAID,
it is still under construction. There are http://www.4raid.com/raidlevels.htm.
some issues that must be handled to make Adaptec, RAID White Papers and
RADD useful. A good thing is that all Benchmark Papers, What Is RAID?,
RAID algorithms work directly in http://www.adaptec.com/worldwide/produ
a distributed environment; however ct/markeditorial.html?sess=no&cat=
distribution generates its own significant %2fTechnolog
problems, which must be handled. Other y
things to consider are the fact that %2fRAID&prodkey=raid_wp_benchmark
algorithms must be able to handle &type=Technology.
unequal amount of storage at each site, it Berkeley, RAID,
must be possible to mix different group http://www.sims.berkeley.edu/courses/is25
sizes in the same network and final, it must 7/f99/Lecture10_257/sld025.htm.
be possible to subtract disk storage from IBM, IBM Hard disk drives | A history of
RADD without the need of a global IBM "firsts" in storage technology | 1973,
reconstruction. There are other http://www.storage.ibm.com/hdd/firsts/n19
technologies that may be alternatives to 73.htm.
RAID. These are for instance: 2DRADD, IBM, IBM Hard disk drives | A history of
C-RAID, RO WB-optimistic and RO WB- IBM "firsts" in storage technology |
pessimistic. Timeline,
http://www.storage.ibm.com/hdd/firsts/tim
10 Conclusion eline.htm.
De Re IT Corp, White Paper: RAID
There are some disadvantages about Systems,
RAID. They are however few compared to http://www.dere.com/en/wp/raid.html.
the advantages, like the high availability Rockwood, Ben, RAID Theory: An
and performance. These are things that Overview, The Cuddletech Veritas
have become more important, especially Volume Manager
since companies to day use Internet for Series CDLTK Cuddletech,
making business on a global market. One http://www.cuddletech.com/veritas/raidthe
must not forget that technology and ory/x26.html, April
techniques are developed and fine 08th 2002.
tuned every day. This means that Solinap, Tom, RAID: An In-Depth Guide
alternative technologies, like RADD, may To RAID Technology,
be a complement to RAID or even replace http://www.slcentral.com/articles/01/1/raid
it. The idea behind RAID is rather simple. /, January 24th 2001.
It is important to understand it, at least a Storagesearch.com, RAID manufacturers
high level. This is important knowledge on STORAGE search.com,
have to be able to make a good decision http://www.storagesearch.com/raid.html.
about what RAID level to use. The choice Stonebraker, Michael, DATA BASE
of level may have great impact of how RESEARCH AT BERKELEY,
well the system will work. Nowadays Department of Electrical
RAID controllers have become cheaper. Engineering and Computer Science
Therefore RAID has become available to University of California Berkeley, CA
a bigger market. It is even possible to find 94720, http://s2kftp.
it in ordinary workstations and small cs.berkeley.edu:8000/postgres/
company servers. papers/sigmodr90-ucb.pdf.
Stonebraker, Michael † - Schloss, Gerhard
A. ‡, DISTRIBUTED RAID -- A NEW
MULTIPLE COPY ALGORITHM, †
EECS Department, CS Division, ‡ Walter
A. Haas
School of Business, University of
California, Berkeley Berkeley, CA 94720,
http://s2kftp.
cs.berkeley.edu:8000/postgres/papers/ERL
-M89-56.pdf.