In this phase, we wish to see
the effect of scheduling policies on the throughput under different transport
- Consider the following network topology.
There is a CBR (Constant
Bit Rate) connection (which uses UDP) from node A to node D sending at a rate
of 3 Mbps. (You may want to set random_
to 1 to get rid of synchronization.) There is another connection from node B to
node D which generates Exponential On/Off traffic (Please refer to part
on Traffic generators in the NS Documentation) over UDP, with an
average burst time of 500 ms and average idle time of 500ms with the burst
rate equal to 6 Mbps. Assume that all packets are 1000 bytes in length.
Provide the graphs for the throughput for both the
connections as a function of time for the following two cases.
scheduler at node C. Assume that link C-D has a drop-tail queue of
length = 40.
Robin Scheduler at node C. Implement the two connections as
belonging to different traffic classes (This sample code
should be helpful). Also, assume drop-tail queues for each class of
length = 20 (each).
For each of the parts above, answer the following question:
- Are the throughputs for the
two connections fair? Why or why not? Please provide a clear reason. You
may attach relevant plots to support your argument.
- We now
consider the topology shown below.
There is an FTP/TCP (TCP Reno) flow from A1 to D1 and another from A2 to
D2. Note that the propagation delay from A1 to D1 is twice as large as the
propagation delay from A2 to D2. Assume that each packet is 1250 bytes in
length. This is the same question as the part 2 in phase 1. The difference
in this case is we wish to implement a Round Robin Scheduler at B. Assume
that the queue for each class at B can hold upto 130 packets. Run the
simulation for 1500 seconds.
provide a graph of the window size as a function of time.
provide a graph of the throughput obtained as a function of time.
Now answer the following questions:
How does the round-robin scheduling compare to
the case of FCFS scheduling at node B?
Does it lead to equal rates (fairness) for the
two flows in spite of a difference in RTTs?
If it is fair, why is it so? If not, why not?
using the trace-all/namtrace-all feature for
long simulations -- it will fill up your disk quota.
all simulations, set window_ variable for all TCP agents to 100000 to ensure
that the receiver window size does not become a bottleneck.
on “Queue Management and Packet Scheduling” and
“Traffic Generators” in the NS Documentation should be useful
for this project.
Submit a report along with the
ns-2 scripts used together in a single tar file. Email the tar file to firstname.lastname@example.org,
with the subject line as "EE122 Project Phase II" by 11.59 pm, April 11, 2006.
- ns-2 is installed on the instructional machines
under /share/instsww/pkg/ns-allinone-2.27-fasttcp. To run it from
your account make sure to
add /share/instsww/ns-allinone-2.27-fasttcp/bin to the PATH
environment variable and /usr/sww/lib to the
- ns-2 online documentation is here.
- gnuplot is a good tool for
generating plots. You will find useful links here.