Network Topologies & Types Of Networks
[edit section] Topologies
Networking topologies are organized by the way in which information "flows" across a network. Below are the basic topologies:
At the core of the Network+ exam and networking concepts in general is the idea of topology, or more specifically, the manner in which data is exchanged over the network. Network topology is mainly a conceptual topic - when we speak of "star" networks or "ring" networks, we are really speaking in terms of the manner in which information is exchanged and not their physical setup. Remember that each topology/network type has its unique advantages and disadvantages that will be tested on the Network+ exam in the form of asking you "which is the most appropriate." Don't try to memorize the perks of each - rather, try to understand the manner in which each allows the exchange of information; then, the advantages and disadvantages will seem only logical to you.
Bus - This is the most simplistic topology in which the nodes of the network are individually linked up to two successive other nodes or another node and a terminating node or terminator. This is considered now an archaic topology, because of the difficulty of troubleshooting network issues (how do you know which node is causing the connection issue?), redundancy issues (if one node fails, the network as a whole can fail), the need for terminators, and the amount of traffic created (every node between A and B must receive the packet that A sends). The nodes linked in this topology are often referred to as "daisy-chained."
Ring - Similar to a bus network in that nodes are linked to each other, but dissimilar in that the ends of a ring network are not terminated because, well, there are no ends! A ring network is something like a "circular" network in which each and every node is linked to two other nodes. This shares many of the same weaknesses as the bus topology, including troubleshooting difficulty, redundancy issues, and traffic created, and also adds an additional difficulty - the difficulty of adding a node to a token ring network.
Star - This is the most typical and practical network setup. In a star network, each node maintains an individual connection to a switch, where all other nodes are connected. Traffic between two known nodes, therefore, only goes through the switch and not through other nodes. This increases the redundancy of the network (one computer faltering will not cause the network to fail), increases data privacy (unicast traffic does not travel through all nodes), and is a relatively easy-to-use setup. Disadvantages include reliance on the switch (a fail-point) and the amount of wiring necessary.
Mesh/"Ad Hoc" - This is a rarely occurring configuration in which every node is connected to every other node; it usually occurs only in wireless networks in "ad hoc" mode, which will be discussed later; in this mode, each wireless card maintains a connection to each other wireless node it wishes to connect with, forming a "mesh" of a network. This is a relatively easy to understand option but is inefficient, requires a large amount of overhead, and is difficult to manage.
Combined or Hybrid - This is simply a topology referring to the case where more than one topology is utilized. For example, you may have three token ring networks connected to a central hub, forming a star of token rings. This is one of many possibilities of a hybrid network.
[edit section] Types of Networks (Access Models)
The Network+ examination is interested in your ability to identify network access models, generally referred to as types of networks. This does not suggest the way in which network nodes are connected or the way that information flows (as do topologies), but rather, the manner and mode in which nodes communicate with each other and share information. There are three basic types:
Decentralized - Often referred to as "peer to peer" network, a decentralized network does not contain any distinctions between client and server. In a decentralized network, every node acts as a client and/or a server depending on the task at hand. For example, many file sharing networks are considered "decentralized" because nodes both download and upload (serve) files. The ease of adding nodes and the ease of setup is a drawing point of decentralized networks, but the pivotal downfall of these networks is their difficulty of maintenance (a setting must be changed on each node to reflect a setting change on the whole network).
Client-Server Access - In this type, nodes can either act as "clients" or "servers," requesting or handing out information. Do not confuse the model with the star topology; though the star topology often utilizes the client-server access model, this does not infer that every client-server network utilizes the star topology. In a client-server network, management is easy and the network can offer services that decentralized networks cannot, but this comes at the expense of difficulty in setup, setup cost, and server reliance.
Centralized - A centralized network is a modified client-server network in which the clients have no individual control; that is, all maintenance and setup occurs at the server level. The extreme ease of management and ability to micromanage is a drawing point of this access method, but the prohibitive cost and inflexibility of the method deter many from employing it.