Mesh topology is a network in which devices are connected to numerous redundant inter-connections, usually between network nodes. Full mesh and partial are two types of these typologies. Its advantage lies in the fact that it is fault tolerant and has the ability to withstand high traffic since data can be transmitted simultaneously from different devices. This is because; nodes can connect with each other disregarding the state of the rest of the network. This means that if one component fails, there is an alternative to it and data transfer loss is not incurred. It is also easy to upgrade and modify without disrupting the nodes. The demerit of mesh topology is that many network connections can cause redundancy, its administration is tough and the overall costs are higher than those of other networks (Stewart and Chapple 2011).
Bus topology is the simplest form of topology. It is characterized by all nodes being connected to one cable through interface connectors. Advantages are: - it is fairly easy to extend and set it up, it requires a smaller length of cable unlike other networks, and its costs are quite low and are efficient in small networks and the LAN. On the other hand, there is a limit to the length of the central cable and also to the number of nodes. The whole network can break down when the central cable develops a problem and it is challenging to trouble shoot a fault at an individual station (Forouzan 2007). The use of terminators is a must and bus network efficiencies decrease, as the devices connected to it increase. It is not efficient in heavy traffic and lastly, its security levels are lower since all connections receive a shared signal from the source.
Ring topologies are connection frameworks on which all nodes are connected to each other, thus forming a closed loop. A token helps in sending and receiving data while each workstation is connected to two others. It is very organized and reduces the chances of collision and traffic flows in one direction at high speeds. Its performance is better compared to bus topologies in the face of traffic. A network server, for controlling connectivity between work stations, is not necessary and additional components do not hamper the network performance. All computers have an equitable resource access. It is slowed down by the process of all packet data movements between all computers and the destination. The whole network can be affected whenever one work station port goes down (Forouzan 2007). The network relies on the wire connecting different components. Network cards and MAU’s used are relatively expensive compared to hubs and Ethernet cards.
In star topology, all components are connected to a hub, which acts as the central device. It is star shaped. It gives better performance as signals are not necessarily transmitted to all workstations. It is also easier to connect new nodes or devices without affecting the rest of the network. It has a centralized management, which helps in monitoring the rest of the network hence easy troubleshooting since failure of a node does not affect the rest of the network. However, if the central device fails, the whole network goes down and the hub increases the overall cost (Stewart and Chapple 2011). The central device also determines the performance of the network and number of nodes connected.
Ethernet is the protocol controlling data transmission over local area network and is most commonly used by technicians. It is very cheap to install and it has higher noise immunity as coaxial cables are well shielded. However, it is difficult to reconfigure once installed and is highly fault intolerant. It is also tricky to trouble shoot and uses special cables that cannot be used for other purposes.
Token rings use special three-byte frames known as, tokens. These tokens usually travel around a ring. Its advantage is in its deterministic nature, where nodes can only transmit at certain well defined times, therefore, minimizing transmission collisions. On the flip side, it requires proper network planning, professional physical installation and it is costly. It is also difficult to trouble shoot. FDDI is based on redundant fiber optic paths through which data travel in reverse directions. It ensures that data is transferable whenever a connection to a “ring” is lost. It supports real time allocation and has a dual fault tolerant ring (Harris 2002). It also, compensates for wiring failures. On the other hand, multiple ring failures are possible, especially as the network grows. Fiber optic cables used are expensive too.
Wireless is a modern approach that uses radio or microwaves in maintaining communication channels between computers. It is flexible, easier to implement and ideal to interior places like mountains. It is also good for temporary network setups. On the flip side, it has a lower speed, less secure, more complex to configure and affected by surrounding.
TCP/IP architecture is much different from the OSI model (Forouzan 2007). TCP is used to verify correct deliveries from clients to servers and aid in the detection of errors. It is also used to check for lost data through the triggering of retransmissions. IP’s on the other hand operates on gateways and is responsible for moving packets of data from one node to another. The agreement of most experts is that TCP/IP has fewer levels. These layers range from between three to five layers. OSI’s on the other hand, have seven layers. A four layered model of TCP/IP is composed of the physical, transport, data link and network layers. The physical layer is used for encoding and signaling, physical data transmissions and hardware specifications. It is also used in topology and design. The data type handled by it is in bits and its scope, is on the electrical or light signals sent between local devices. It uses CSMA/CD for accessing purposes.
Data link, which is the second layer, is used for data framing, addressing, error detections, media access controls, logical thinking and defining requirements of the physical layer. It handles data types in terms of frames and its scope is on low level data messages in local devices. The common technologies and protocols it handles are: - the Ethernet Family, IEEE 802.11 (WLAN, Wi-Fi), IEEE 802.2 LLC, Token Ring, FDDI and CDDI; Home PNA, Home RF, ATM, SLIP and PPP. The network, which forms the third layer, is responsible for Error Handling and Diagnostics, Logical Addressing, Encapsulation, Routing, Fragmentation and Reassembly. It handles data in the form of datagram / Packets and its scope is on messages between local/ remote devices. The common protocols and technologies it utilizes are; IP, IP NAT, IP sec, ICMP, Mobile IP, IPX, DLC, PLP, IPv6 and Routing protocols such as RIP and BGP. The transport layer is used for Process-Level Addressing, Multiplexing/Demultiplexing, Segmentation and Reassembly, Connections, Flow Control, Acknowledgments and Retransmissions. It transmits data in form of data grams/ segments. Its scope is on communication between software processes and the protocols and technologies used are; SPX, NetBEUI/NBF, TCP and UDP.
Sockets on the other hand are Sub-routine packages that provide access to TCP/IP on most systems. The OSI model consists of seven layers. Apart from previously mentioned matter, it also comprises of the session, presentation and application. These last three layers are commonly referred to as, the top layer. The session concerns itself with conference establishment, management and termination. It transmits session data and its scope is on sessions between local or remote devices (Forouzan 2007). The protocols and technologies used are: - NetBIOS, Sockets, Named Pipes and RPC. The presentation layer concerns itself with data presentation, encryption and compression. The data transmitted is, the encoded user data and its scope is on the application data representations. The protocols and technologies used are: - MIME, SSL, Shells and Redirectors. The last layer, application, is concerned with user application service. The data transmitted is user data and its scope is on application data. The protocols and technologies it utilizes are: - DNS, NFS, HTTP, Teln, RMON, BOOTP, DHCP, SNMP, FTP, POP3, TFTP, SMTP, IMAP and NNTP.
An organization may decide to employ the use of wireless LAN as it is convenient since users can access resources offered by the network from a location within the networking environment. It is also important in organizations in terms of mobility as it can move from conference rooms to offices with laptops without loosing signals. It can increase productivity due to convenience as the worker can work from any location (Stewart and Chapple 2011). It is easier to set up because of fewer cables and leaves an organization with less work, for instance, if it is a mobile organization that requires moving all the time. Furthermore, it’s easier to expand its area of coverage and its hardware is at a modest cost. This helps organizations to streamline on their expenses.
Wireless are not good enterprise networks because of their security hiccups. A network can easily be hacked and information that’s crucial stolen. In a wired network, however, an adversary would first have to overcome the hurdle of tapping into the network and avoiding detection by the server. The range of the wireless network is usually in terms of tens of meters and it would be insufficient for a large structure (Harris 2002). It is also not reliable because of interferences, for example, by walls. The speed is also limiting as a result of the interferences and it the normal range is between 1- 108 Mbit/s unlike wired networks which are between 100 Mbit to several Gbit/s.