|
(a) Architecture, Algorithm and Implementation
are the three aspects of computer networks. Briefly explain the following
concepts, and identify to which aspect of networks they refer.
1. Shortest Path Routing [1]
2. EIA and RS232C [1]
3. TCP/IP for HP-UNIX [1]
4. ATM Network Model [1]
5. Congestion Control [1]
6. TCP/IP for Linux [1]
1. Algorithm
2. Architecture
3. Implementation
4. Architecture
5. Algorithm
6. Implementation
(1 mark each, provided some explanation is given. If the aspects are
simply listed, award up to 2 marks over all)
(b) Identify the main differences between the two Internet transport
layer protocols TCP and UDP. [3]
TCP offers a session-based connection; UDP
sends single messages
only or: TCP is connection-oriented; UDP is connection-less (1 mark)
UDP transfers a limited amount of data; TCP is unbounded (1 mark)
TCP gives reliable transmission (via retransmission if necessary);
UDP is unreliable (1 mark)
Give credit for other valid differences.
(c) Although hosts on the Internet are identified using IP addresses,
most network applications do not use IP addresses directly. Briefly
explain what they do instead, and how this scheme is implemented.
[3]
- The application refers to a host using
a domain name instead of an IP
address (1 mark)
- The name to address translation data may be stored on each
computer or on the DNS domain name server (1 mark)
- The user application will send domain name to server, and server
will
send back the IP address (1 mark)
Give credit for valid alternative explanations.
(d) Briefly describe three levels of
hierarchy at which routers are used. [3]
- At the lowest level in the hierarchy the
routers are used to connect
multiple sub networks within a LAN (1 mark)
- Routers are also used to connected different networks within the
same domain Autonomous systems. (1 mark)
- At the highest level in the hierarchy routers connect networks from
different domain or Autonomous systems. (1 mark)
Do not award marks simply for listing layers.
(e) For each of the following examples
of Internet Control Message Protocol
messages, identify to which class of ICMP message they belong to.
[3]
-Echo request & reply
-TTL expired or bad port number
-Routing redirect
-Network management
-Error
-Control (1 mark each)
(f) Explain two reasons for using layered
protocols, using examples to illustrate
your answer. [4]
- Layered protocols allow the breaking up
of the design problem into
smaller and manageable pieces (1 mark) Example: application
designer does not need to understand ethernet protocol (1 mark).
- Layering means that protocol can be changed without affecting higher
or lower layers (1 mark). Example: TCP service parameters can be
tuned without affecting basic functionality of application layer (1
mark).
Give marks for valid alternative answers, especially in the area of
examples.
(g) Describe two advantages and two
disadvantages of having international
standards for network protocols. [4]
Advantages:
- If everyone uses the standard, everyone can talk to everyone else,
that is provides easy interconnectivity (1 mark)
- Widespread use of any standard will give it economies of scale,
such
as VLSI chips (1 mark)
Disadvantages:
- Frequent standardisation may lead to poor standards (1 mark)
- Once a standard has been widely adopted, it is very difficult to
change
it, even it new and better techniques or methods are discovered (1
mark)
Give marks for other valid comments.
(h) IPv6 is gradually replacing the
familiar IPv4. Explain two major differences in IPv6, giving reasons
why the changes have been made. [4]
- IPv6 uses 128 bit addresses, instead of
the current 32 (1 mark). This
is because, due to poor address-space utilization, the world is
running out of 32-bit addresses (1 mark).
- IPv6 will allow simplified autoconfiguration, machines generating
their
own addresses when connected to a network (1 mark), at present the
DHCP offers only some of this functionality (1 mark).
Give full credit for valid alternative answers.
|