Computer Networking • CS 514

Test 7, 10 December 2014

---

This test has six questions. The test runs the entire class period, starting at 4:30 and ending at 6:20. When you finish the test, you may leave.

Always justify answers unless the question explicitly states it’s not necessary. Some questions require reasonable assumptions in order to be answered. In such cases the answer should clearly state the assumptions made. Answers should be no longer than 75 to 100 words.

1. Determine as precisely as possible the word or phrase being defined by the following:

   • An agreed-upon bit pattern serving as the divisor in calculations.

   • The frame-forwarding structure defined by bridge forwarding tables.

   • The amount of time a channel spends transmitting a signal divided by the total time.

   • A (fully-qualified domain name, IP address) pair.

   • Translates a bit stream into a bit stream more suitable for transmission.

   ---

   Keeping in mind that giving an example of X is not defining X, briefly (in one or two sentences, 25 to 30 words) define:

   • An agreed-upon bit pattern serving as the divisor in calculations:

   The generator polygon (not cyclic redundancy check, which contains a generator polygon, but also contains much else).

   • The frame-forwarding structure defined by bridge forwarding tables:

   A spanning tree.

   • The amount of time a channel spends transmitting a signal divided by the total time:

   The effective transmission rate, or the utilization.

   • A (fully-qualified domain name, IP address) pair:

   A DNS A record for IPv4 or an AAAA record for IPv6.

   • Translates a bit stream into a bit stream more suitable for transmission:

   A line encoding.

2. Describe how you would modify the HTTP application-level protocol to use out-of-band control.

   ---

   Following FTP's lead, to establish a connection with out-of-band control, HTTP would need two connections, one for data and the other for control. The control connection lasts the life of the connection, while the data connection can be created as needed or created once at connection set-up and kept for the life of the connection. HTTP can use a single continuously connected data channel because HTTP response messages are sized delimited and don't need an explicit end signal.

3. Suppose the TCP transport-layer protocol used stop-and-wait for flow control. Describe how the stop-and-wait TCP would provide congestion control.

   ---

   TCP congestion control is based on a non-trivial sliding window. Stop-and-wait has a trivial sliding window; using stop-and-wait in TCP breaks congestion control. Assuming congestion appears to the end-points as dropped packets, under normal operation stop-and-wait has at most one packet (data or ack) in flight at any time. If congestion is detected, stop-and-wait has two options: drop packets into the network less often, or drop smaller packets into the network (assuming acks are already as small as possible, the data packets would be smaller).

4. A four-port, layer-three router is part of a network having eight-bit addresses. The router's forwarding table contains:

   		output
   prefix		port
   01		3
   100		3
   101		2
   1101		0
   *		1

   Assuming the router forwarding algorithm uses maximal prefix matching, what are the address ranges sent to each output port? Justify your answer.

   ---

   Addresses in the range 00000000₂ through 00111111₂ match no prefix and are sent to the default port 1.

   Addresses in the range 01000000₂ through 01111111₂ match prefix 01 and are sent to port 3.

   Addresses in the range 10000000₂ through 10011111₂ match prefix 100 and are sent to port 3.

   Addresses in the range 10100000₂ through 10111111₂ match prefix 101 and are sent to port 2.

   Addresses in the range 11000000₂ through 11001111₂ match no prefix and are sent to the default port 1.

   Addresses in the range 11010000₂ through 11011111₂ match prefix 1101 and are sent to port 0.

   Addresses in the range 11100000₂ through 11111111₂ match no prefix and are sent to the default port 1.

5. L_B is a set of LANs inteconnected by bridges and L_H is identical to L_B except the bridges in L_B are hubs in L_H. Which of the LANs L_B and L_H has the higher aggregate bandwidth? Justify your answer.

   ---

   Assuming there are at least two constituent LANs in the network (otherwise why have bridges or hubs in the first place), L_B has a higher aggregate bandwidth than does L_H. A bridge preserves the constituent LANs’ collision domains, which means transmissions can occur simultaneously on each constituent LAN. Hubs, on the other hand, don't preserve collision domains, and only one host at a time on the combined LAN can transmit at a time.

6. Red light has a wave length of 700 nm; violet light has a wave length of 400 nm. What is the bandwidth of the visible spectrum?

   ---

   This question is lightly tricky because “bandwidth” was given without units. Assuming units in Hz, the bandwidth is f_v - f_r = 0.32 PHz, where

   fv	= (c m/sec)/(400·10-9 m/cyc)
   	= (300·106 m/sec)/(400·10-9 m/cyc)
   	= (0.75·1015 cyc/sec)

   and

   fr	= (c m/sec)/(400·10-9 m/cyc)
   	= (300·106 m/sec)/(700·10-9 m/cyc)
   	= (0.43·1015 cyc/sec)

   If, on the other hand, the units are in bits/sec, a few more assumptions are needed. First, assume an errorless channel, which enables Nyquist's equation. Second, assume a two-symbol alphabet, which makes the log equal to 1. The bandwidth is 0.32 Pbits/sec.