Itbm Networking Architectire
Itbm Networking Architectire
NETWORK ARCHITECTURE
Learning Objectives:
After studying this chapter, you will be able to:
Understand the importance of network architecture in data communication.
Understand different types of network implementation.
Understand different types of communication medium and their applications.
Differentiate Internet, Intranet and Extranet.
amiliar with different network terminologies.
1
A !etwork is a group of systems that are connected to allow sharing of resources such as files or
printers, or sharing of ser"ices such as internet. #he physical connection between the
systems$de"ices in a network is established using either cable media or wireless media. !etwork
de"ices that originate, route and terminate the data are called network nodes. A client is a de"ice
that sends re%uest to the ser"er of the network and a ser"er is a computer system$de"ice that
recei"es the re%uest, processes it, and returns the re%uested information back to the client.
!etwork architecture is a framework of the network&s physical components and their functional
organi'ation and configuration, its operational principles and procedures, as well as data
formats used in its operation. (ased on the framework there are two types of network
architectures:
Centralize Arc!itect"re
#istrib"te Arc!itect"re
6$% Centralize Arc!itect"re
#he first computers were large, expensi"e, and difficult to manage. #he computer executed one
)ob at a time. #erminals, which came later, pro"ided the user with a new mechanism to interact
with the centrali'ed computer. #hese terminals, howe"er, were merely input$output de"ices that
had no independent processing power. All processing took place on the central computer* hence
it is named as the centrali'ed computing. !etworks, therefore, ser"ed little purpose other than to
deli"er commands to and get results from the powerful centrali'ed system. #o this day, large
mainframe systems are still being operated around the world, most often by go"ernments and
large corporations. #hese early computing models worked well in large organi'ations that could
)ustify the need for these expensi"e computing de"ices. +owe"er, one of the drawbacks was that
the mainframes were not flexible in their placement ,some were the si'e of a large room- and did
not scale down to meet the needs of smaller organi'ations. !ew ways of sharing information
were necessary to allow computing power to be shared efficiently on smaller networks.
A centralize net&'r( is a network in which most communications are routed from one ma)or
central hub$system. In the centrali'ed network computing model, the clients with low or no
processing capabilities use the resources of a high.capacity ser"er to process information. #he
clients only connect to the ser"er and not to each other. igure /.0 shows a centrali'ed network
computing model. A client in need of ser"ices sends the re%uest for the same to the ser"er which
processes the re%uest and responds back to the client with the re%uired ser"ice. It offers
greater security because all information processing is controlled in a central location. In addition,
if one terminal breaks down, a user can go to another terminal and all of the files will still be
accessible. Depending on the system, they may e"en be able to resume their session from the
point they left. Another benefit of centrali'ed networks is the ease of maintaining accurately
updated lists of data that can be easily accessed from all points.
2
)ig"re 6$%: an e*a+,le '- centralize net&'r( arc!itect"re
#raditionally, this type of networking was only found in Enterprise 1e"el (usinesses. In recent
time, due to reduced ser"er and network costs of centrali'ed network architecture, this type of
computing deployed in many smaller and medium si'ed businesses.
C!aracteristics '- a Centralize Net&'r(
A centrali'ed system consists of a large data center that hosts all ser"er resources.
2oftware upgrades can be done from a centrali'ed location.
#he data center incorporates power.insulating de"ices such as Uninterruptible 3ower
2upply ,U32- and 4hot.sites5 or 4cold.sites5 contingencies.
(usiness re%uirements associated with reducing cost and security re%uirements are usually the
dri"ing forces behind centrali'ed networks. #here are mainly two ways of centrali'ing network
implementation i.e. 6lient ser"er model and +osted client model.
Client./erver 0'el
1et us consider figure. /.0 which is a client.ser"er model. #he clients need different
ser"ices for which they send re%uests to the central hub or ser"er. #he ser"er upon
recei"ing the re%uest processes the re%uest and responds accordingly to the client. 2o we
may say that a 4client5 is a computer system that initiates communication with the ser"er
to a"ail some resources. A 1ser"er5 is the one that shares its resources with client
3
Workstation
computer. #his sharing of resources is also known as time.sharing because multiple
applications are allowed to share the same resources at the same time. #his
communication between the clients and the ser"er always follows a re%uest.response
networking pattern thus making it compulsory for the computers to ha"e a common
language and common rules of communication, defining as 4communications protocol5.
)ig"re 6$2: Re3"est.Res,'nse net&'r(ing bet&een a client an t!e server
Client /erver Envir'n+ent
#his en"ironment is comprised of a ser"er which pro"ides system resources
,processing power, storage, software, applications, ser"ices etc- to se"eral connected
clients.
#he ser"er controls what a user has access to. (oth software and data can be
accessed by any of the clients. #he ser"er contains all software. e.g.. If one wants to
upgrade the most recent "ersion of 7ord, he has to do it once on the ser"er and then
e"eryone can use it. 1ikewise with operating system or internet software etc can also
be shared.
#he ser"er can hold and protect all data.
H'ste.Client 0'el
Another method of centrali'ed computing is hosted.client model. In this model,
processing and storage on powerful ser"er hardware located in a data center, rather than
in a local office. 8rgani'ations are relie"ed of many responsibilities in owning and
maintaining an information technology system.
A,,licati'ns '- Centralize Net&'r(s
AT0s: A#9 machine is an example of centrali'ed networking with which e"eryone is
familiar in using. A#9s function as terminals. All processing is done on the mainframe
computer to which the A#9s are connected.
P"blic Instant.0essengers: 9ost of the 3ublic Instant.9essenger platforms use the
centrali'ed network architecture to pro"ide ser"ices to the users.
4''gle Cl'" C'+,"ting: :oogle 6loud computing is also used for employees5
centrali'ed networks.
Avantages
Ad"antages of centrali'ed network architecture are:
4
CLIENT SERVER
NETWO
RK
Centralize ata +anage+ent: In a centrali'ed network, data is stored on the ser"er for
increasing the reliability of data because all data modifications are stored at a central
location.
Hig! level '- sec"rit5: #he centrali'ed network computing model is a highly secured
network model because network security can be implemented and monitored centrally
from the ser"er.
C'st e--ectiveness: #he o"erall cost of setting up a centrali'ed architecture is "ery low.
Li+itati'ns
6entrali'ed network architecture has the following limitations:
L'& ,er-'r+ance an net&'r( s,ee: A single ser"er manages numerous re%uests
simultaneously for increasing network traffic and conse%uently reducing the speed for the
performance of the network.
Central ,'int '- -ail"re: #he ser"er is the central and only place for storing data and
processing all client re%uests. If the ser"er fails, the whole network collapses.
I+,'rtant C'nsierati'ns
A centrali'ed system is considered only if prere%uisites are the following:
#ata center !ar&are c'sts #he cost of installing high.end ser"ers and clusters in the
data center against the administrati"e cost sa"ings of centrali'ing the ser"ers must be
weighed. It is recommended that the back.end ser"ers are clustered to build high
a"ailability and redundancy into the system, but this in"ol"es greater costs up front.
+owe"er, these costs may be more than offset by reductions in operational costs,
infrastructure costs, reduced downtime, and greater scalability.
C'ntingenc5 ,lanning 7hen you ser"er and data resources across the organi'ation are
centrali'ed, the possible single points of failure are increased. 6ontingency plans must be
formulated.
#ata st'rage 7ith larger centrali'ed data "olumes, more reliable storage systems must
be used to impro"e the integrity of data. In addition, by reducing the complexity of ser"er
infrastructure, one can more readily restore ser"ices and data when a failure occurs.
/ec"rit5 A centrali'ed model gi"es easier security management, thus it gi"es a greater
degree of control. #his control makes it easier for security staff to maintain up.to.date
"irus signatures and take timely action in response to security incidents. Another
ad"antage of a centrali'ed design is that it locates your ser"ers in a data center which you
can physically secure.
6$2 #istrib"te Arc!itect"re
5
Distributed 6omputing is a trend in modern day business en"ironments. #his is the opposite of
centrali'ed computing, which was pre"alent during the early days of computing. Distributed
computing is the allocation of resources both in hardware and software to each indi"idual
workstation or office location. In contrast, centrali'ed computing exists when the ma)ority of
functions are carried out or obtained from a remote centrali'ed location. #he distributed network
computing model allows all network computers to take part in processing but at their respecti"e
ends separately. #his model allows sharing data and ser"ices but does not help the other network
computers in processing. In this network model, a processing.intensi"e task is broken into a
subset of tasks and distributed among multiple nodes. #he nodes work on their indi"idual subsets
of tasks. #he following figure ,figure /.;- shows the distributed network computing model:
)ig"re 6$6: A istrib"te net&'r( arc!itect"re
A distributed computer system has many benefits o"er a con"entional centrali'ed network.
Desktop computers ha"e ad"anced so rapidly that their potential performance far exceeds the
re%uirements of most business applications. #his results in most desktop computers remaining
idle ,in relation to their full potential-. A distributed system can utili'e the potential of these
systems to maximi'e efficiency. +owe"er, it is debatable whether these networks increase o"erall
effecti"eness. All computers ha"e to be updated indi"idually with new software, unlike a
centrali'ed computer system. Distributed systems still enable file sharing and all computers can
share peripherals such as printers and scanners as well as modems, allowing all the computers in
the network to connect to the internet.
A collection of distributed computers systems are components of a larger computer network, held
together by local stations of e%ual importance and capability. #hese systems are capable of
running independently each other.
C!aracteristics '- a #istrib"te Net&'r(ing /5ste+:
6
CENTRAL
LOCATION
INTERNET
A branch office or distributed networking deployment is one where numerous branch offices or
smaller distributed sites ha"e slow connections to a corporate hub or data center. #he branches
contain their own exchange ser"ers, domain controllers, and global catalog ser"ers. A distributed
networking system is usually adopted when the network cannot handle traffic to a central hub for
the ser"ices. 2o the operating system and networking ser"ers are placed locally. User
re%uirements may be another factor. If the re%uirements for user experience and a"ailability
cannot meet by connecting to a data center, you may ha"e no choice but to place ser"ers in the
remote sites. A distributed computing system deployment has the following characteristics:
#he branch office locations usually contain a small or "arying number of users.
#he network connections between the branch office locations and the central hub or data
center are typically low.bandwidth, high.latency, or unreliable.
#he main reasons behind deploying a distributed networking system include the following:
#he company&s network infrastructure cannot handle traffic to a central hub for ser"ices.
#he user re%uirements dictate that a ser"er be placed locally to pro"ide optimal user
experience and a"ailability.
A,,licati'ns '- #istrib"te net&'r(s
Telec'++"nicati'n net&'r(s
#elephone networks and cellular networks
6omputer networks such as the Internet
7ireless sensor networks
Net&'r( a,,licati'ns
7orld wide web and peer.to.peer networks
9assi"ely multiplayer online games and "irtual reality communities
Distributed databases and distributed database management systems
!etwork file systems
Distributed information processing systems such as banking systems and airline
reser"ation systems
7
Real.ti+e ,r'cess c'ntr'l
Aircraft control systems
Industrial control systems
Avantages '- #istrib"te Net&'r( Arc!itect"re
/calabilit5: Enterprise solutions that rely on a single enterprise ser"er ine"itably suffer
from performance issues as the enterprise grows and the ser"er is o"erwhelmed.
9oreo"er, single ser"er solutions are highly susceptible to network failures.
E--icienc5: 2ecurity managers control the flow of data and decision.making. 1ocal data
and decisions can be transmitted to each indi"idual site, minimi'ing network bandwidth,
and allowing global managers to focus on truly global issues. At the same time, centrally
located global managers can easily run reports, make changes, and "iew the status of
local sites without needing to login to multiple separate systems.
C'st: 2er"ers and software at each local site can be appropriately si'ed to meet the
specific needs of each site, without re%uiring installing an expensi"e ser"er at e"en the
smallest sites.
Reliabilit5: Distributed !etwork Architecture is much more tolerant of network and
hardware failures than a single ser"er approach.
I+,'rtant C'nsierati'ns
#ata st'rage: 7ith distributed ser"ers, the ser"ice infrastructure is more complex,
which makes it more difficult to restore ser"ices and data when a failure occurs.