The differences between a MAN and a LAN are:
- MANs cover wider geographical areas. For example, MANs have the capability to cover an entire city like Chicago. LANs can only cover an entire building.
- MANs recover more quickly to switch/router failures.
- Most MAN's topologies are based on a ring both physically and logically.
- MAN have the ability to allocate bandwidth on demand.
All MAN networks are based on two different kinds of technology: SONET (Synchronous Time Division Multiplexing Technique) and Ethernet. SONET's have a low failover rate because of their topology. SONETs are consisted of a ring topology, but these rings are composed of multiple rings within the ring that help with backup in case of a segment failure. However, SONETs have their disadvantages. This technology can be expensive because of its complexity and they have issues with scaling from T-1 transmissions to 1-Mbps, 10-Mbps, and 100-Mbps.
Ethernet MANs are less expensive and are better understood and can have the ability to scale from 10-Mbps to as much as 10 Gbps. The problem with ethernets, however, is their failover rate. Routing/Switching failures are higher than the SONET technology.
Wide Area Network Basics:
A wide area network is "a collection of computers and computer-related equipment interconnected to perform a given function or functions, typically using local and long-distance telecommunications systems." WANs connect each workstation through the use of a "mesh" design and use a routring device to transport data across the network. In this case, neighbors will only be connected to other neighbors. All WANs are collections of two different types of equipment: nodes and workstations. Nodes are devices that allow other workstations to access the physical network and are also transfer points for passing data through the network. To support the wide area network, a network cloud is provided within the WAN. The network cloud is the subnetwork of the WAN, which will contain a collection of nodes and interconnecting telecommunication links. The network cloud is the underlying physical interconnection of nodes and communications lines that transfer data from one location to another. The cloud is responsible for guiding the data to the proper destination node. The network cloud is simply a vehicle for the data that is to be transported from one workstation to another.
Types of network clouds:
There are three different types of network clouds: circuit-switched network, packet-switched network, and broadcast.
Circuit-switched network is a network cloud in which a dedicated circuit is established between the sender and the receiver, and all other data will pass over this dedicated line. When a connection is requested for this kind of network, time is needed to establish the circuit and to tear down the circuit. The problems with circuit-switched networks is they only contain one connection, and the circuit may not be utilizing the connection fully due to the sporadic behavior of circuit-switched networks.
Packet-switched network will transmit all data messages using fixed-sized packages. The connection is not established. The data will be broken down into various packets from the sender and then reassembled at the requested destination. There are two types of packet-switched networks: datagram and the virtual circuit. Datagrams allows each data packet to follow its own unique path from each node, which will help with congestion or failures. The problem of datagrams is that each packet will need to analyzed. This could lead to inefficiencies and wasted time. Virtual circuits will send out a notification to routers that will participate in the transportation of various data packets through the network. Once each router has completed its task, that dedicate line will dissolve. This is why it is called a virtual circuit because a dedicated line is established, but that line is then removed when the message has arrived to its specified destination.
Connection oriented versus connectionless network applications:
Connection-oriented network applications provide "guarantees" that the data that is transferred will not be lost over the network. The data will also be sent to the intended receiver. This type of network will provide a "reliable service," which will need a logical connection to be established between two endpoints.
Connectionless network application does not require an established logical connection between two endpoints. This leads to the loss of data and information over the network.
Routing:
Weighted network graph is a way to assign conditions for various connections between various devices. This will help traverse the network. Each communication link will have a weight which will contain a specific meaning. One line connection will have a weight will different meanings. Some will have dollar costs for using the connection and others will have time-delay costs.
Dijkstra's least-cost algorithm is an algorithm that will calculate the least-cost path through a network. The algorithm is executed and then stored within the node that has executed the code. Some nodes will also share the algorithm. However, this algorithm is time consuming. Therefore, the execution of the code is only periodic or when something changes within the network.
Flooding is when a node will take an incoming packet from another node and will transmit the packet to every outgoing link. There are two common rules associated with this technique:
- The node receiving the packet will not send a copy back to the sender.
- A "hop limit" will be required to limit the number of copies that are created.
Centralized routing is routing information that is stored in a central location. When a router needs information for various incoming packets, the router will "query" this central location for the information that is associated with the packet.
Distributed routing uses an algorithm to generate routing information and dictates that this information to be stored at various distributed locations within the network.
No comments:
Post a Comment