Tutorial #13
- What does CSMA/CD mean? CSMA/CD is sometimes referred to as the "polite
dinner table" algorithm? Can you think of a reason for this?
- Traditional Ethernet/802.3 is a shared medium,
multi-access network technology.
- What do each of these terms mean?
- Does this pose any potential security risks? Explain. What about a
modern network based on a switching hub?
- What is interesting about the Ethernet/802.3 "MAC address"? Discuss.
- In the lecture, it was stated that an Ethernet/802.3 collision occurs when
two stations start to transmit at the same time. Discuss the significance of
the phrase at the same time as used in this context.
- The 10baseT configuration has captured the Ethernet market from thin wire.
Discuss reasons why network planners and managers might prefer this
technology. There are lots of reasons.
- Switching Hubs build a table which maps destination MAC
addresses to physical interface port numbers. How do they build this table?
- An IP packet is "encapsulated" into an Ethernet frame for delivery within
the network. How will this encapsulation differ in the case where the packet
is sent to a router for "Internet Delivery", compared to IP "local delivery"?
- In the
lecture, it was stated that the key difference between Ethernet and IEEE
802.3 LANs is the meaning of the 16 bit "type" field (used as a "length" field
in 802.3). In many LANS, Ethernet and 802.3 frames co-exist perfectly happily.
How can a receiving station know whether to treat the field as a
length or as a type?
- Assume[1]
a one megabyte file must be transferred across a network. Ignoring delays
caused by waiting for access and other overhead (ie, counting only the data
transferred), how long would it take to send the file across an Ethernet?
Across a Fast Ethernet?
- The ARP protocol maps IP addresses to MAC addresses. Can
you think of an alternative approach to providing these mappings? Why is ARP
better?
- What would you expect the destination MAC address to be in an ARP request?
- Engineering research question[2]:
A shared-medium (non-switched) Ethernet is generally regarded as heavily
loaded (approaching overloaded, in fact) if the network utilisation goes over
(approximately) 20%. This, on the face of it, seems a low value. What do you
think is going on?
- Opinion question: ATM is probably the dominant high-speed
networking technology at present. What do you think is the particular
attraction of ATM networks over other high-speed technologies, on the basis of
the (minimal) material presented in the lecture?
- Serious Research question #1:In the lecture, it was mentioned that Gigabit
Ethernet is compatible with 10/100Mbps Ethernet. What does
this mean? It's OK to guess...
- Serious Research question #2: What is the historical origin of the
difference between the frame formats of Ethernet and IEEE 802.3?
[1] From Comer,
Computer Networks and Internets 2/e P91 (paraphrased).
[2] You are not expected to understand this.
Copyright © 2004 by Philip
Scott, La Trobe University.