All slide sets and reading assignments (in homeworks for the 4000
level students) will be part of the exam. Here are a list of
topics/concepts which we will expect you to be prepared in. 

The chapters/sections will be the same as what applied to exam 1 and
exam 2. In addition, for the link layer the sections are mentioned
below. I have also put up an extra set of notes on 



Introduction/Basics: Chapter 1
===================

Connectivity: direct, indirect, point-to-point, multiple access,
	      switched, inter-network
How to build Scalable connectivity: filtering, forwarding, switches,
		routers, hubs...
Internetworking issues: heterogeneity, scale
System design concepts: tradeoffs in resources, 
	Multiplexing: statistical multiplexing, stability, tradeoffs
		      in delay/throughput
	Circuit switching vs packet switching
Protocols, abstractions, layering, service models,
service/peer-to-peer interfaces, encapsulation, design perspectives of
users, designers, providers. 


From the text: 
	Network edge, core, access networks, physical media.

Application layer: 
=================

- Principles, design of HTTP, FTP, SMTP, DNS, Sockets
- How these protocols reduce to a request/response primitive and are
  thus mapped onto TCP or UDP depending upon their requirements.


Transport Layer:
===============

- Services and principles
- Multiplexing, demultiplexing.
- Connectionless vs connection oriented: understanding of issues.
- UDP
- Principles of reliable data transfer, redundancy models (temporal),
solid understanding of mechanisms like checksums, acks/naks, sequence
numbers, windows, and how they can be combined to give different
degrees of reliability 
- Go-back-N and selective repeat: the tradeoffs and design issues
- TCP: services, segment structure, connection management, reliable
  data transfer, delayed acks, flow control, RTT estimation and
  timeout setting algorithms. 
- Principles of congestion control and TCP congestion control:
  self-clocking (a.k.a ack-clocking), dynamic windows, slow start,
  congestion avoidance, AIMD and fairness analysis, 
- Latency modeling: understanding of core analysis and implications
  regarding transfer times of small/medium/large packets   

Network Layer:
=============
- Internetworking problems: heterogeneity and scale
	- Heterogeneity: overlay and translation models. Implications
	  of overlay model: address and packet format
	  mapping/resolution
	- IP forwarding paradigm: use of address structure to deduce
	  direct/indirect connectivity, Layer 3 forwarding table, no
	  broadcast at the network layer, Use of any available layer 2
	  forwarding to reach destination or next hop
	- Getting a datagram from source to destination.

- IP addressing structure, address allocation issues:
	- Impact of address structure and understanding of why layer 2
	  does not scale without it (because it has a flat structure).
	- Class-ful model (class A, B, C, D, E), dotted decimal
	  notation, why it led to inefficient address allocation 
	- subnet masking and CIDR (supernet masking) and DHCP.
	- Implications on forwarding/routing algorithms:  use of masks
	  in forwarding and routing
	- Address allocation in CIDR: hierarchical allocation,
	  variable length prefixes depending upon size of networks. 
	

- What IP does not provide? How is this consistent with the end-to-end
  principle ? How is this different from the telephony model ?
 
- Mapping problems: IP fragmentation/reassembly, address resolution
(ARP), path MTU discovery

Routing:
=======
- intra-domain and inter-domain routing
- principles of routing: consistency, completeness, scalability
- link-state vs distance vector approaches 
[note: I have also put up some additional notes on IP and routing on
 my backup web page]
- Dijsktra and Distance vector (bellman-ford) algorithms. 
- RIP and OSPF


Modeling:
=========
        (1) probability mass functions and distribution functions
        (2) some of the most common distribution: binomial, exponential,
                geometric and Poisson
        (3) M/M/1, M/M/1/K and M/G/1 queues: concentrate on which model
                to use rather than trying to solve for the Markov chains
                etc. 
        (4) Little's theorem and the relationship between E(t) and E(n)

Datalink Layer (Chapter 5, sections 5.1 to 5.7)

        (1) Services offered by the datalink layer
        (2) Why do need to provide reliability in the datalink when
                we already have it in network layer?
        (3) Error detection and correction
                - parity checks
                - CRC
            you should concentrate on how to generate parity bits and
            CRC checksum. Wether they are error detection or correction
            codes.
        (4) You should understand the difference between the 3 types of
            channel access schemes: fixed (TDMA, FDMA etc), random
            (Aloha, CSMA/CD etc) and taking turn (polling etc). What
            are the advantages and disadvantages of one over the other
            and under what circumstances would you use them.
        (5) Details of Aloha, slotted Aloha, CSMA/CD, Ethernet and 802.11
        (6) Difference between hubs, bridges and switches. Advantages
            and disadvantages of each. You should understand the self
            learning properties of bridges and what are spanning trees.