1. TCP waits until it has received three duplicate ACK before performing a fast retransmit. Why do you think the TCP designers chose not to perform a fast retransmit after the first duplicate ACK for a segment is received?

1

• 1. TCP waits until it has received three
  duplicate ACK before performing a fast
  retransmit. Why do you think the TCP designers
  chose not to perform a fast retransmit after the
  first duplicate ACK for a segment is received?

The designer probably felt that waiting for two
subsequent packets (rather than 1) was the right
tradeoff between triggering a quick
retransmission when needed, but not
retransmitting prematurely in the face of packet
reordering.
2

• Refer to the figure that illustrates the
  convergence of TCPs additive increase,
  multiplicative decease algorithm. Suppose that
  instead of a multiplicative decrease, TCP
  decreased the window size by a constant amount.
  Would the resulting additive increase additive
  decrease converge to an equal share algorithm?
  Justify your answer using a diagram similar to
  the figure.

equal bandwidth share
R
loss decrease window by factor of 2
congestion avoidance additive increase
Connection 2 throughput
loss decrease window by factor of 2
congestion avoidance additive increase
Connection 1 throughput
R
3

• Refer to the following figure. In Figure
  (a), the ratio of the linear decrease on loss
  connection 1 and connection2 is the same as
  ratio of the linear increases unity. In this
  case the throughput never move off of the AB line
  segment. In Figure (b) the ratio of the linear
  decrease on loss between connection 1 and
  connection2 is 21. That is, whenever there is a
  loss, connection 1 decreases its window by twice
  the amount of connection 2. We see that
  eventually, after enough losses, and subsequent
  increases, that connection1s throughput will go
  to 0, and the full link bandwidth will be
  allocated to connection 2.

4

• 3. Recall the idealized model for the
  steady-state dynamics of TCP. In the period of
  time from when the connections window size
  varies from (WMSS)/2 to WMSS, only one packet
  is lost (at the very end of the period).
• a. Show that the loss rate is equal to
• L loss rate 1/(3/8w23/4w)
• b. Use the above result to show that if a
• connection has loss rate L, then its
  average
• bandwidth is approximately given by
• Avg. BW of connection 1.22MSS/RTTsqrt(L)

5

• Consider the following network. With the
  indicated link costs, use Dijkstras shortest
  path algorithm to compute the shortest path from
  F to all network nodes.

2
14
4
1
9
1
6
2
3
1
1
4
1
3
6
Dijsktras Algorithm
1 Initialization 2 N A 3 for all
nodes v 4 if v adjacent to A 5 then
D(v) c(A,v) 6 else D(v) infty 7 8
Loop 9 find w not in N such that D(w) is a
minimum 10 add w to N 11 update D(v) for
all v adjacent to w and not in N 12 D(v)
min( D(v), D(w) c(w,v) ) 13 / new cost
to v is either old cost to v or known 14
shortest path cost to w plus cost from w to v /
15 until all nodes in N
7
6
8

• Recall the two FEC schemes for Internet phone.
  Suppose that the first scheme generates a
  redundant chunk for every four original chunks.
  Suppose the second scheme uses a low-bit-rate
  encoding whose transmission rate is 25 of the
  transmission rate of the nominal stream.
• a. How much additional bandwidth does each
  scheme require? How much playback delay does
  each scheme add?
• b. How do the two schemes perform if the
  first packet is lost in every group of five
  packets? Which scheme will have better audio
  quality?
• c. How do the two schemes perform if the
  first packet is lost in every group of two
  packets? Which scheme will have better audio
  quality?

9

• 6. What would be preemptive priority queueing?
  Does preemptive priority queueing make sense for
  computer networks?
• 7. Give an example of scheduling discipline that
  is not work conserving.
• 8. What are some of the difficulties associated
  with the Intserv model and per flow reservation
  of resources?

10

• 9. A connection has a mean rate of 1 Mbps, a peak
  rate of 10 Mbps, and a delay jitter of 500 ms.
  What is the amount of buffer needed at the
  receiver to remove the delay jitter?
• 10. Compute the max-min fair allocation for
  sources A, B, C, D and E, when their demands are
  2, 3, 4, 4, 5, and the resource size is 15.
• 11. Connections A and B are continuously
  backlogged during time 0,1 and have weights 1
  and 4. A receivers 4 Kbits of service in 0,1.
  What service is B guaranteed to receive with FCFS
  and GPS disciplines?

11

• 12. Packets of length 100 and 200 bits from
  connections A and B arrive at an empty FQ
  scheduler at time 0. If the line rate is 100
  bps, (a) at what real time do the packets
  complete service? (b) what is the corresponding
  round number (virtual time) when each packet
  completes service? (c) If a packet of length 10
  arrives on connection A at real time 1.5 s, what
  would be its finish number?
• 13. If connections A and B in the above have
  weights of 2 and 5, respectively, recompute parts
  (a), (b) and (c).