A group of interconnected computers is referred to as a computer network. Networks can be classified according to various characteristics (Tanenbaum, 2002). First, scale, where the main type of networks are Personal Area Network (PAN), Local Area Network (LAN), Metropolitan Area Network (MAN) and Wide Area Network(WAN). The categories basically depend on the area of coverage. Secondly, Connection Method, which is a classification based on the hardware used in the interconnection of individual devices on the network (Tanenbaum, 2002). Under this there are Ethernet, Optic fibre, Wireless and HomePNA networks.
Thirdly, functional relationship, which is a classification method based on the interaction between the elements of the network under which there are Active, Client-server and peer to peer networks (Tanenbaum, 2002).
A local area network (LAN) is a computer network that links a group of computers in close proximity for instance in office buildings or at home. This network is often connected with other LANs, Internet or a Wide Area Network (WAN). Most WAN’s are built with relatively less costly materials such as Ethernet cables, network adapter and a hub. There is also the possibility of having a wireless LAN. There are specialized operating system softwares that could be used in the configuration of the LAN for example Ms Windows provide Internet Connection Sharing (ICS), a software that supports IP controlled access to LAN resources (Tanenbaum, 2002). The most common LAN is the Ethernet LAN.
A LAN could be composed of two computers and up to a thousand computers. An Internet Protocol (IP) which is a type A LAN can in theory accommodate up to 16 million computers organized into logical group known as subnets. The physical implementation of a network or the topology plays a great role in ensuring that the network satisfy its role in resource sharing, mobility of data, reliability, replication and other network uses.
The topologies which include bus, star, mesh and hybrid are varied in their implementation and thus have varying advantages and their suitability to an organization would greatly determine their efficiency.
Local Area Networks
Based on the above classification there are a number of aways in which a LAN can be implemented and thus there is need to look keenly at the methods, their advantages, requirement and disadvantages.
a) Wired LAN
This is a LAN in which the devices are interconnected by the use of either thin or thick coaxial cables (Tanenbaum, 2002). The transmission of signals between devices are governed by the Institute of Electrical and Electronics Engineers (IEEE) standards. Ethernet Protocols refer to the family of LAN covered in the IEEE 802.3 standards. Under this standard there are two modes of operation namely half duplex and full duplex. The simplex mode of transmission allow data to be transmitted in only one direction. In the half duplex mode data is transmitted through the Carrier-Sense Multiple Access/Collision Detection(CSMA/CD) protocols in a shared media.
The main disadvantages of this mode of transmission, which allow transmission in only one direction at any given time, are inefficiency and distance limitations. It is important to note that the link distance in this mechanism is limited by the Media Access Control (MAC) frame size. The modification implemented on this mechanism to minimize the disadvantage is carrier extension which ensures that a minimum of 512 bytes is transmitted in any given transmission. The full duplex mode allows communication to and from the communicating devices to pass through a communication media at one given time and is thus more efficient, though the implementation is more tedious relative to half-duplex.
The Ethernet system consists of three elements: The physical medium used in signal transmission, a set of protocols embedded on the Ethernet interface to facilitate the interaction between different network devices and the Ethernet frame which consists of standardized bits. Like in all IEEE 802 protocols, the data-link layer is divided into the MAC and MAC- client sub-layers. The primary responsibilities of the MAC sub-layer are data encapsulation and assembly, frame parsing before transmission and transmission management which include recovery in case of failure. The MAC-client sublayer may be regarded as the Logical Control Link (LLC) which provide the interface between the MAC and other protocol stacks or bridge entity which provide interfaces between LANs.
In this mechanism each computer operates independent of other computers as there are no central controllers. All devices are attached to the media, which is a shared transmission system. To send data a station first listens to the network and then transmits only when the channel is idle. At any given time all stations have an equally opportunity to transmit data, though access to the channel is determined by the MAC mechanism. This system is based on CSMA/CD.
All devices look at the destination address embedded on the frame and delivery initiated if the data matches the current destination, otherwise the frame is released. The logical topology implemented is a single channel that carries the signal to all stations. It is possible to link multiple segments, to form a larger LAN using repeaters, though such growth does not have a specific root. Each segment must never be connected in a loop, since the implementation of that will lead to failure in operation. Even though the media may be physical connected in a star pattern attached to a repeater, the logical channel is basically bus.
In large networks there is use of bridges and repeaters that are used to amplify the signals so that they may be transmitted over long distances. The implementation of such devices are stipulated by the Ethernet Standard used. The main advantage of the Ethernet LAN as a wired network are:
The hardware involved in the interconnection, which include hubs and switches, are quite inexpensive.
There is increased reliability associated with the use of cables, hubs and switches.
Wired LANs offer better performance than the wireless LANs.
The disadvantages in using a wired LAN are:
A Security risks as wired Ethernet hubs do not support firewalls.
In convenience in terms of dealing with wires that are often bulky and give the organization an untidy look.
a) Optic fibre LAN
This is a form of cabling uses fibre optic (Tanenbaum, 2002). It also implements the IEEE protocols. The cable is expensive and quite difficult to install. This mechanism is mainly used to interconnect buildings or equipments that are too remote to be connected using the typical Ethernet cables. It is referred to as the 10 base F Ethernet protocol though it can be implemented by other networking systems that require high speeds.
A 10 base F link is implemented in a point to point link and is not a shared link as in 10B2 cabling systems. This cabling system is common between repeaters of a bridge between two LANs. The main advantage of this cabling system are:
High immunity to disturbance.
Ease of upgrade
The disadvantages of using optic fibre LANs are:
The optic fibres which is the main transmission media is expensive.
Segment length in the later versions is less than 2 Km.
External transceiver is required.
It is important to note that the network diameter in any LAN must not exceed 5Km.
A LAN that does not rely on wired physical connection is referred to as wireless local area network (WLAN) (Tanenbaum, 2002). A WLAN can therefore be an extension or alternative to an existing wired network. It is important as it adds to network flexibility to the network as it allows the ease of movement while still being connected. The WLANs transfer data at rates that vary from 1 to 54 Mbps, though some manufacturers can offer up to 106 Mbps. The mechanism used in transmission of signals in a WLAN is broadcast and thus the signal is available to all that have the equipment to receive it within the broadcast network. There is need to put security mechanism in place to ensure security of the data in transmission. A WLAN signal can cover a small office or a large network though the most common provide access within a radius of 300 feet. The equipment that are used in the installation of a WLAN are readily available in most retail stores. The hardware in this category generally subscribe to the 802.11 Ethernet standard and WPA2 standards due to security concerns.
The enterprise class WLAN uses a large number of access points in broadcasting the signals. The features of access points in comparison to those of home WLANs are: Better security, authentication tools for integration and remote management. The access point are designed to work together so that they can cover a wide area. The hardware employed by this class adhere to the 802.11a, b, g Ethernet standards though they do also adhere to WPA2 and 802.1 standards due to security requirements. The main advantages of this approach are:
Data mobility is increased.
Highly convenient as the inconveniences and untidiness associated with the use of cables and fibres is eliminated.
The increased mobility could lead to increased performance by employees in an organization as they can work from any where within the broadcast network.
The deployment of a wireless LAN requires little apart from a single access point. Therefore the deployment is quite simple relative to wired networks that have the additional inconveniences of having to deal with cables.
The main disadvantages are:
The method is costly in terms of initial installation and the implementation of protocols that would ensure the security of the data within the broadcast region.
The range of coverage of the wireless networks is quite small and therefore its application is unsuited for large organizations though the addition of repeaters and access points may increase the network coverage.
The speed of the wireless LAN is quite low if compared to that of wired networks.
a) HomePNA LAN
HomePNA is an alliance that develops home networking solutions for the distribution of signals over the existing coaxial cables and phone lines. The alliance also has internationally recognised open and interoperable standards that are aimed at ensuring best practices. HomePNA enables users to meet their multimedia demands such as IPTV and Voice Over Internet Protocol (VOIP) through Quality of Services (QoS) (Tanenbaum, 2002). There are several advantages and disadvantages that associated with the implementation of the above in a LAN. The advantages are:
No special wiring or device required.
Existing services are not disrupted by the implementation of Home PNA as it operates at different frequencies.
High data rates that makes it able to carry data of multimedia data of any format and size.
Up to 64 devices can be connected at any given instance, this is a large number in a home setting.
The network coverage allows for up to 300M between two adjacent devices in a communication session.
Uses standard Ethernet wires which are quite cheap.
The QoS mechanism eliminates any collisions that are associated with Ethernet mechanisms.
Very flexible as technologies like Wi-Fi can easily be added to the home networks.
The major disadvantages associated with HomePNA LANs are:
The network is limited to only 64 computers and thus not suitable for large companies.
The cost benefit ratio associated with the use of HomePNA is low for large offices that transmit ordinary data.
It does not coexist with some network platforms for instance DOCSI
The available chipsets are very few.
2. Functional Relationships
a) Active LAN
An active LAN is a network that does not operate in a fixed manner. The schemes involved in the operation of such a network are quite dynamic, fully programmable an provide interfaces for execution of programs received from other nodes via the network (Tanenbaum, 2002). An active network does not necessarily mean a static topology. To enable the use of active LAN, there is need to use component based software engineering, code mobility( Use of Java) between end hosts and field programmable gate arrays that will aid high speed packet processing. There are two models of active LANs . The first model runs the code on data flowing through the switches, this model is characterized by the separation of programs and packet demultiplexing to the right programs. This model is referred to as discrete active LAN. Secondly, an integrated active LAN in which individuals can inject programs into the LAN. The integrated model is characterized by restricted programming languages and carriage of code by data packets. The programming of the nodes can be done by either the user or network administrator. The main motivation behind Active LANs is the leverage of computation in the LAN. Other advantages involved in using active LANs include:
Automation of adaptive streaming.
Data aggregation which has the average effect of reduced data volumes.
Reduction of latency.
Provides the end user with customized services.
Increased safety in the execution environment.
There are several disadvantages involved in this approach the most common are:
Increased expenses in terms of programming knowledge.
Security threat in that malicious programs could be retrieved from other devices tat may harm the host if run.
There is need for larger bandwidth if the interactive nature of the network is to be maintained at high level.
b) Client-server LAN
This architecture distinguishes clients from servers. A client is defined as a node seeking a resource, while the server is defined as a resource provider. Most application use a client server model especially the web browsers. A client-server LAN therefore consist one or more server machines from whom files and shared programs reside (Rodriguez, 1995). The LAN servers are usually big ( in terms of storage space an disk space) and fast as they have to meet the network requirements. The clients on the other hand are meant to serve one user at a time and therefore have lower specifications. It is possible in a client server architecture for a machine to provide specialized services, an example is the for print server that is charged with the responsibility of printing requirements of the network. Advantages of Client-Server LAN are:
Centralized management make it possible to implement security mechanism.
A fast server implies ease of access to resources, furthermore the location of resources can easily be identified as they reside in one computer.
The disadvantage to this approach include:
The centralization of resources could lead to a big loss in case of server failure.
Poor server speeds could translate to poor access speed to resources by the clients.
c) Peer to Peer LAN
A peer to peer (P2P) computer networks uses diverse connectivity between devices in a network and cumulative rather than conventional bandwidth and resource use. P2P is used in connecting nodes via ad hoc connection. A P2P LAN is therefore important in sharing data that is in digital data and multimedia data.
A pure P2P does not implement the client-server notion (Tanenbaum, 2002). The implementation is in a way that the peers are equally charged with the client and server responsibilities. This differs from the traditional client-server mechanisms in that communication is not always from a specific source. The concept of P2P is evolving and is increasingly being used in dynamic active and distributed networks. An important objective in the peer to peer network is that all clients provide resources, which are: Bandwidth, storage space and computing power. Therefore, as the number of nodes increase there is an associated increase in the total capacity. This aspect is not true in the client sever architecture where there is fixed resource based at the server, and therefore more data imply slow data transfer rates to all users.
A distributed P2P network increase reliability as data can be retrieved from other sources in case of failure in the primary source. This is due to the replication of data in a distributed network and therefore the elimination of single points of failure.
The network in a P2P is made up of the nodes and the links between them. The P2P networks can be categorized broadly into two namely: Structured and unstructured P2P (Mitchell, 2008). An unstructured P2P is characterised by the arbitrary establishment of overlay links. They are generally easy to construct as new peer only needs to copy existing links from other peers to form its own links. In this system when a device seeks data in the network, the query has to go through a number of peers that have shared data. The disadvantage of this method lies in the fact that the resolution of queries is not guaranteed as rare data is almost impossible to find as there is no correlation between a peer and the content it manages. Flooding also causes traffic which if accumulated to large level is undesirable as it has a negative effect on the network efficiency.
The structured P2P employs a protocol that is consistent over the network thus ensuring that any node can efficiently locate some desired file no matter how rare the file is. This calls for a more structured approach to the overlay links. The most common of the P2P networks is the distributed hash tables (DHT), in which hashing is used in the assignment of file ownership to a particular file. The advantages of peer to peer LAN are:
There is reduced cost in cabling as computers only have to be connected to other computers. This is not the case in the client-server architecture where the connection to the server by any computer in the network is mandatory.
The implementation of P2P is quite simple and so is the technique used in transmission.
Elimination of risks associated with centralized storage.
The disadvantages are:
Management of resources and security measures are quite complicated as there are several points of entry.
The link to the network and reliability in terms of finding resources is quite low especially in an unstructured P2P.
The choice of the mode of implementation of LAN is directly affected by several factors notably: The size of the network, the nature of the environment and the available resources. Therefore, before choosing the kind of LAN one should implement, there are several factors that must be addressed. The organization or home need to ask itself the following questions:
What is the nature of the data involved in transmission?
What is the available resources allocated to the LAN and its management?
What is the size of the organisations?
Should the network be interactive?
Does the organization and the end user have the knowledge needed to implement a specific LAN?
What is the efficiency of the network type and is it reliable?
The answer to the above questions greatly affects the choice of the systems to be implemented. The efficiency of the network therefore greatly depends on the nature of the organization, network type, topology implemented and the management of the network resources.
Mitchell, B. (2008). Local Area Network. Retrieved on 19 Jul 2008 from http://compnetworking.about.com/cs/lanvlanwan/g/bldef_lan.htm.
Rodriguez, C.(1995). The LAN Connection. Retrieved 23 Jul 2008 available from http://www.uic.edu/depts/accc/newsletter/adn005/lancon.html.
Tanenbaum, A. S.(2002).Computer Networks. Indiana, Prentice Hall PTR