As said in the previous lecture, the data link layer has 2 purposes:
We will focus on the second purpose.
Collision is the interference of transmissions from multiple parties that use some medium at the same time. We can detect collisions because error detection indicates error. If there is a collision, you need to retransmit (unless error correction can repair).
In this lecture we will try to answer the following questions: How do we reduce collisions? Can we completely avoid collisions and how? Can we detect collisions before the error detection step and how? How do Ethernet and Wireless work? How do local area networks work?
Channel Allocation Protocols can be divided in various categories:
Static channel allocation. Every party always get the same share of the medium.
For example, there is always 5 parties and each of them always get a 100ms to sent. When they are done, the next person comes and so on.
Static protocols are rarely applicable. In most cases a dynamic approach is preferred.
Channel allocation protocols can also be divided into:
Dynamic channel allocation. The opposite, each party get a different share of the medium.
Without Carrier-Sense Multiple Access (CSMA). The protocol can’t/doesn’t sense if the medium is in use. We’re “blind” of who is using the network.
With Carrier-Sense Multiple Access. The protocol senses if the medium is in use.
These protocols are pretty simple (as we can’t know anything about what is going on in the network).