5.Routing with a Distance Vector Protocol
Choosing the right physical topology allows a company to expand its networked serviceswithout losing reliability and efficiency. Network designers base their topology decisions uponthe enterprise requirements for performance and reliability. The star and mesh topologies arenormally deployed in enterprise environments.
One popular physical topology is the star. The center of the star corresponds tothe top of the hierarchy, which could be the corporate headquarters or head office. Branchoffices at multiple locations connect to the center, or hub, of the star.A star topology providescentralized control of the network. All crucial services and technical staff can be located in oneplace. Star topologies are scalable. Adding a new branch office simply requires one moreconnection to the central point of the star. If an office adds several branches to its territory,each branch office can connect to a center hub in its own area, which then connects back to themain central point at the central office. In this way, a simple star can grow into an extendedstar, with smaller stars radiating out from the main branch offices.The star and extended startopologies create a single point of failure. Mesh topologies eliminate this problem.
Each additional link provides an alternate pathway for data and addsreliability to the network. With the addition of links, the topology becomes a mesh of interconnected nodes. Each additional link adds cost and overhead. It also adds to thecomplexity of managing the network.
Adding redundant links only to a specific area of an enterprise creates a partialmesh. This topology meets uptime and reliability requirements for critical areas like serverfarms and SANs, while minimizing additional expenses. The other areas of the network are stillvulnerable to failures. Therefore, it is essential to place the mesh where it provides the mostbenefit.
When no downtime is acceptable, the network requires a full mesh. Each node in afull mesh topology connects to every other node in the enterprise. This is the most failure-proof topology, but it is also the most expensive to implement.The Internet is an excellent example of a meshed network. Devices on the Internet are notunder the control of any one individual or organization. As a result, the topology of the Internetis constantly changing, with some links going down and others coming online. Redundantconnections balance the traffic and ensure that there is a reliable path to thedestination.Enterprise networks face some of the same issues as the Internet. Therefore,processes are put in place that allow devices to adapt to these constantly changing conditionsand reroute traffic as appropriate.