CHAPITRE 3: Basic Computer Networking

1 INTRODUCTION
Early chapter, we have seen that the types of network i.e. LAN, or WAN
are establish for sharing data, to provide services, to allow for administration and
security and to reduce equipment cost. To achieve this centralized, Distributed
and collaborative systems are use for computing of data. Actual Network
implementation can be done by using peer-to-peer, or server based networks.
Your Network can provide services like File, Print, Application and database etc.
Transmission media is a path way network entities use to contact each other.
Computer transmission media includes cables and wireless technologies that
allow network devices to contact each other. To reduce their design complexity
most networks are organized as a series of layers or levels. The Rules and
conventions used in this convention are collectively known as layer protocol.
Basically a protocol is an agreement between the communicating parties on how
communication is to proceed. A set of layers and protocol is called as Network
architecture. A list of protocol used by a certain system, one protocol per layer is
called as protocol.
3.2 ORGANIZATIONAL COMPUTATIONAL MODELS
Whether a LAN or WAN, the overall goals of network are to establish a
means of sharing data, to provide services, to allow for administration and
security, and to reduce equipment cost. Three models, or methods of organization,
are available for networking.
1. Centralized All processing is done at one location
2. Distributed Independent operation and local task
3. Collaborative Computers cooperate and share the load
3.2.1 Centralized computing
Centralized computing was the first method of networking implemented.
As the name implies, all networking is done at one central location. The best
example of this would be a UNIX host with a number of dumb terminals. The
dumb terminals are nothing more than input/output interface into the host, and all
processing actually takes place at the host. Because all interaction is at one
location, all the terminals directly connected to the host and never connect with
each other.
Whole processing of data will take place on centralize machine, but
because of this system client’s machine has to send all data to central node, which
will increase unnecessary traffic between server and client machine. As central
machine has to respond each and every node speed of this system is low.

Centralize computing system

Merits of Centralize System
• Excellent security
• Centralize administration as both application logic and data resides on the
same machine
Demerits of centralize system
* Mainframe computers are very expensive to buy, lease, maintain and use.
 • The imitation is that both the application arid database live within the
same machine process thereby offering no way to truly partition the application
logic beyond the physical limitations of the mainframe.
3.2.2 Distributed Computing
With distributed computing, the dumb terminals are replaced PCs. The PCs
can function separately and also interact with servers. Task are run locally, and
data is exchanged, but without the server's performing any direction. A good,
example of this scenario would be an NT server acting as file server with a
number of Windows98 clients are capable of independent operations. The
windows 98 clients are capable of independent operation. When they need to
perform a task involving a file, they obtain it from server and perform the
operation they need. The server gives them the file but doesn't tell them what to
do with the data that was requested.
In this system application logic was executed of the client workstation
instead of the server. These servers also provided access to computing resources
like printers and large hard drives
Merits of distributed computing-
• Low-cost entry point with flexible arrangement
• Computer resources can be added or reduced as and when necessary,
using this system.
Demerits of distributed computing
• As central administration is not there this will provide share level security.
• As client machine can do processing, client's machine need large amount
of power to run the application.
 Taking into account the demerits of centralize system and distributed
system architecture, collaborative computing architecture made its advent.
3.2.3 Collaborative Computing
Collaborative computers is also known as cooperative computing, enables
computers to not only share resources (such as files) but also share processing.
There are two methods by which this can invoke. A server might borrow an entire
processor from an idle machine to perform an action, or the server might share
part of processing with client.
A classic example of this environment is Microsoft SQL server. When a
client requests data, SQL server does some of the processing an sends data to the
client for the completion of processing on that system. In all cases, the software
must be written to take the advantage of absence of such software.

3.3 DIFFERENCE BETWEEN CENTRALIZE DISTRIBUTED AND

COLLABORATIVE SYSTEM
Case study
Let us assume that we have a set of data stored in a database file namely
student _info.mdb (Microsoft access file). This file holds the details of the marks
stored in different -subjects by the students in their public examination. A client
may want to know as to how many students have scored 100 percent in more than
two subjects. A query is sent-to obtain the results. We shall discuss about the
processing of the distributed and collaborative systems when a query to obtain
the results satisfying the above-mentioned, criteria is issued.
1. Query sent to centralize system
In centralize type of computing server is a main, component all others
are dump terminate or just input output nodes. Student_info database resides
on centralize machine. Client's machine not hava any processing power. so
query is sent to the server and server will do all the processing and processed
results are sent to the client' machine.
2. Query sent to distributed system
In this case, the logic of query, is processed and evaluated at the client
machine itself. The query logic realize that it needs to access a table namely
student info in the MDB in order to process the request. Hence it requests the
student info table with all rows across the network before it applies the
conditional clauses, which specifies the criteria that client is looking for.
So, when SQL statement is used against a MOB, it is processed by the
client machine and only a file I/O request is sent across the network to retrieve
the required data in the from of disk blocks. No logic is executed at the server
end except the transferring of file disk blocks. This is just a distributed
computing.
3. Query sent to collaborative system
In collaborative architecture the actual SQL statement is sent across the
network and processed by an application running locally on the server
machine. As the SQL statement is processed on the server, only results need
to send back to the client. This is a vast improvement over the distributed
system. The query looking for student's detail having scored 100 % in two or
more subjects is evaluated at the server end and only those records satisfying
these criteria would be passed over the network instead of all records of the
table.
Thus after receiving records from server, client's machine can perform
rest of his work and display records satisfying condition to the user.

3.4 NETWORKING MODELS

For actual network implementation we can use following networking
models
1. Peer-to-Peer Cheap to implement, minimal security
2. Server-based Requires a dedicated server and good security
3.4.1 PEER-TO-PEER
 In a Peer-to-Peer network you take the machine currently in existence,
install networking cards in them, and connect them through some type of
cabling. Each machine is known as Peer and can participate in the sharing of
files or resources. No server is required, so there is no additional cost for a
dedicated machine, but there is also no real security.
Peer-to-Peer networks require an operating system that can understand
networking and function in this (Peer-to-Peer) way. Microsoft Windows 95,
Microsoft Windows 98, Windows NT server and Windows NT workstation
can all function in Peer to-Peer environment.
If file and print sharing has been enabled on a Windows 95 system, for
example, you can create a share by selecting a folder and choosing to share it.
By default, no password is associated with it but you can choose to assign one
that a user must know in order to access the resource. Access permission can
be Read-Only, Full or depend on password this is known as share level
security. Access is gained when a user supplies the correct password to access
the share.
Peer-to-Peer networking works in small environments. If you grow
beyond approximately 10 machines, the administrative overhead of
establishing shares, coupled with the lack of tight security, creates a
nightmare.
Advantages of peer-to-peer network
• Server is not required
• No additional cost for dedicated-machine
Disadvantages of peer-to peer network
• Provides share level security
• Can work in small environments only.

3.4.2 SERVER BASED

In the presence of server, be it on NetWare Or NT, you can implement
user level security on your network. With the user level security, permissions
are based on how the user logged on and was authenticated by the server.
Every user has an account. In this environment, you can assign permissions to
shared based on user permissions or group permissions. In short you must have
server on the network in order to have user level security, but you can have
share level security with or without server. This scenario also known as
client/server networks (explain previously in chapter 1), server-based
networking's down side is that it requires a dedicated machine (the server); the
upside is that you gain centralize administration, you can add ail users at one
location, control logon scripts and backups; and so on. With centralized
authentication, you can identify a user to your entire network based on his
logon name and password, not based on each share he attempts to access.
Advantages of Server based network
• Provides user level security
• You always gain centralize administration
• Can work in big environments also
Disadvantages of server-based network
• Dedicated machine is required
• Cost of the system is more compared to peer-to peer networks.
-. Peer-to-Peer networks can exist comfortably within server-based
networks. In many business combinations of two models are used. A server-
based network is used to provide e-mail; and other resources to all users, and
Peer-to-Peer networks are established within divisions to share resource
among select users. Microsoft, also calls Pee-to-Peer networks workgroup and
server- based networks domains. These terms are used interchangeably in
almost air Microsoft documentation.