The construction of computer networks can be traced to the need to share the resources of powerful and expensive computers (mainframes). The technology of the networks, and further the emergence of personal computers at low cost have allowed revolutionary developments in the organisation of computing resources.
You can indicate at least three strengths of a network of computers compared to traditional mainframe. There is fault tolerance as the failure of a machine does not block the entire network and you can easily replace faulty computers (the cheap components and a company can afford to keep spare parts in stock). There is also the cost angle, because as I say, computer hardware and software costs less than those for mainframes. There is also the issue of gradual growth and flexibility (scalability) because adding new capabilities to an existing network and its expansion are simple and inexpensive.
A network has some weaknesses compared to a mainframe such as security. An attacker can gain access more easily to a computer network than to a mainframe to the limit of just being able to physically access the network cabling. Furthermore, once a worm has infected a network system, it spreads rapidly to all others and the work of disinfection is very long, difficult and unclear in its completion. There is also the issue of high maintenance costs. With the passage of time and updates, and with the addition of new functions and services, the network structure tends to expand and become more complex and computers that are part of it become more heterogeneous, making maintenance more expensive in terms of working hours. Beyond a certain limit of size of the network (about 50 computers) becomes necessary to perform hardware and software updates on an entire group of computers rather than on individual machines, nullifying the advantage of low cost hardware.
There is a wide variety of network technologies and organisational models, which can be classified according to different aspects. It is something you will have to consider when you create your data centre design.
There is classification based on the geographical which depends on the geographic extent and distinguishes between different types of networks. It comes to personal network, or PAN (Personal Area Network) if the network extends around the user within several meters. It comes to LAN or LAN (Local Area Network) if the network is extended within a building or a district with an area within a few miles/Km. It comes to wireless network or WLAN (Wireless Local Area Network), if the local network is based on a technology in radio frequency (RF), allowing mobility within the coverage area, usually around hundred meters outdoors. It comes to the university network or CAN (Campus Area Network), meaning the internal network to a university campus or at a set of adjacent buildings typically separated by land owned by the same entity which may be connected with their cables without using the services of telecommunication operators. This condition facilitates the realization of a network of interconnected high performance and low cost. You are talking about underground network or MAN (Metropolitan Area Network) if the network is extended within a city. It comes to wide area network or WAN (Wide Area Network) if the network extends beyond the limits indicated above.
About Author : Emma Hamilton is a freelance content writer and she writes articles and blogs on data centre design and others server related technologies.