CS/ECE 438: Communication Networks for Computers Spring 2007
Problem Set 6 Due: start of class, Wednesday, April 18th
Internetworking and End-to-End Protocols
Assigned reading: Peterson and Davie: Chapters 4, 5 and 9.1. All problems carry equal weight. For full credit, show
all work.
1. TCP RTT Estimation
One difficulty with the original TCP SRTT estimator is the choice of an initial value. In the absence of any special
knowledge of network conditions, the typical approach is to pick an arbitrary value, such as 3 seconds, and hope this
will converge quickly to an accurate value. If this estimate is too small, TCP will perform unnecessary
retransmissions. If it is too large, TCP will wait a long time before retransmitting if the first segment is lost. Also,
the convergence might be slow.
a. Choose = 0.8 and SRTT(0) = 3 seconds, and assume all measured RTT values = 1 second with no packet
loss. What is SRTT(19)? Recall, SRTT(k + 1) = * SRTT(k) + (1 - ) * RTT(k + 1).
b. Now let SRTT(0) = 1 second and assume RTT values = 3 seconds and no packet loss. What is SRTT(19)?
2. TCP Slow Start
Although slow start with congestion avoidance is an effective technique for coping with congestion, it can result in
long recovery times in high-speed networks.
a. Assume a roundtrip time delay of 75 ms (about what might occur across a continent) and a link with an
available bandwidth of 1 Gbps and a segment size of 576 octets. Determine the window size needed to
keep the pipe full and the time it will take to reach that window size after a time out using the Jacobson
Algorithm.
b. Repeat for a segment size of 16 Kbytes.
3. Domain Name Service (DNS)
a. DNS uses UDP instead of TCP. If a DNS packet is lost, there is no automatic recovery. Does this cause
a problem, and if so, how is it solved?
b. In addition to being subject to loss, UDP packets have a maximum length, potentially as low as 576
bytes. What happens when a DNS name to be looked up exceeds this length? Can it be sent in two
packets?
c. Can a machine with a single DNS name have multiple IP addresses? How could this occur?
d. Can a computer have two DNS names that fall in different top-level domains? If so, give a plausible
example. If not, explain why not.
4. Adaptive Retransmission, Part I
a. What is the retransmission ambiguity problem addressed by the Karn-Partridge algorithm? How does
the algorithm avoid the ambiguity?
b. Write a C program to estimate the frequency of unnecessary retransmission with the original TCP
adaptive retransmission algorithm. In particular, write a one-million-iteration loop that calls a function
double get_rtt ();
to measure RTT, calculates an estimated RTT, and counts the number of times that the next measured
RTT exceeds the current estimate. Use a damping factor of 0.15 (α on P&D p. 390 is 0.825) and
double-precision floating-point variables. Programs longer than 50 lines will receive no credit.
c. Repeat part (b) for Jacobson’s algorithm (Jacobson-Karels in P&D, defined on p. 392), using δ = 0.15.
