EE3A2 Computer Networking
EE3A2 Computer Networking is a 3rd year undergraduate course at The University of Birmingham, UK.
The lecture slides and content will be subject to minor updates prior to presentations.
Course Text
The recommended course text is the 2nd edition Communication Networks by Albert
Leon-Garcia and Indra Widjaja, McGraw-Hill.
The book website http://highered.mcgraw-hill.com/sites/007246352x/ contains selected "hints and answers" and copies of the figures.
See the CANVAS course page for erratas for editions 1 and 2.
Course Resources
Laboratory Activity Notes
Lecture Slides and Tutorial Questions
Miscellaneous
Minor Notes and Corrections for the 2nd Edition of the Course Text
- pg368. A note. "In broadcast networks a single transmission medium is shared by a community of users. For this reason, we also refer to theses networks as multiple access networks." CSMA is indeed an abbreviation of Carrier Sense Medium Access. But "multiple access" is an ambiguous descriptor that does not discriminate between broadcast and non-broadcast networks. Non-broadcast networks can also be "multiple access" networks (Wikipedia Non-broadcast Multiple Access Network).
-
pg 428. A correction. End of first paragraph: "A minislot time defines a time duration that is at least as big as two propagation delays." Then in 2nd paragraph: "When allowances are made for four repeaters, the delay translates into a maximum end-to-end propagation delay of 51.2 microseconds." The end-to-end propagation delay is tprop, If the "minislot time" is at least 2tprop then it is at least 102.4 microseconds. However, the commonly quoted Ethernet slot time is 51.2 microseconds. I believe the confusion of the factor of 2 is due either to a muddling of end-to-end delay (tprop) and round-trip delay (2tprop) or because there is a factor of 2 engineering tolerance between the round-trip delay caused by the physical maximum distance of the network and the slot time. If one estimates the length of the maxium network based on frames of 512bits and a signal propagation of about 2x10^5 km/s then one finds that the Ethernet standard allows a network of about half this size. If there were no engineering tolerance, there would be no allowance for delays introduced by repeaters and no allowance for a segment being a centrimetre longer than it should.
- pg523. A correction. "The Bellman-Ford algorithm (also called the Ford-Fulkerson algorithm) .." This is not correct. Lester R. Ford worked on both algorithms. The Ford Fulkerson algorithm computes maximum flow in a network. The Bellman-Ford algorithm computes shortest paths. The algorithms are distinct.
>