TABLE OF CONTENTS Computer networking relies on sharing of limited network resources for communications. If communications resources were unlimited, a direct point-to-point link of unlimited bandwidth would be possible between any two end-points that wished to communicate. Clearly, this is not possible in the real world. Therefore, in the real world, two end-points that wish to communicate, unless they are connected to the same LAN, need to communicate through a network of links that have the following characteristics: (1) the links are shared between traffic from multiple sources to multiple destinations, (2) the links have limited bandwidth, (3) the links have propagation latency (i.e., time taken for data to traverse the links), and (4) the nodes directly connected to each link have limited processing power and memory capacity.
How should these limited resources be shared? The easiest way is to treat all traffic the same, meaning all packets are treated the same, accordingly to loose general principles of fairness. Thus, all packets are processed according to the order in which they arrive, in a first in first out (FIFO) manner. If so many packets arrive at a node that its buffers are filled, subsequently arriving packets may be dropped. The Internet, like many other networks, behaves like this.
However, this simple treatment of packets is not sufficient for today's multi-service networks, which support a variety of services and applications. The different applications have different requirements on the communications services provided by the network. For example, voice and video have...
