Transcript Wasim & Hernan

CS603 - Wireless Communications &
Networks
Summer II 2003
Project Progress Presentation
GENERATION OF NS2 TRACE FILES
FOR SATELLITE-WIRELESS NETWORKS
Wasim Mahmoud El-Hajj
Hernan H. Aranda Gomez
What’s the problem?
 NS2 doesn’t support the implementation of both satellite and wireless
networks at the same time. Satellite links differ from wireless links in the
following respects:
1. Transmit and receive interfaces must be connected to different
channels.
2. There is no ARP implementation, which indicates that there is no
hierarchical addressing.
3. There is no radio propagation model.
 To attach number of wireless nodes to a base station, we have to specify
IP addresses to each of the nodes and associate them with the base
station. This is done using hierarchical addressing by dividing the nodes
into domains and clusters. So if a wireless node wants to send a packet
to another node outside its domain, it has to send the packet to its base
station and the base station will route it to the particular destination node.
If the destination node was part of a satellite network, then NS2 will fail to
deliver the packet because of former reason (# 2 above).
What’s our goal?
 The problem can be fixed by extending the NS2 code in order to make
the satellite network support hierarchical addressing and radio
propagation.
 Also, we can take the two separate trace files and try to combine them
using different approaches, defined later, to form a new trace file that
shows the two networks communicating together.
 We are going to implement the second approach and generate a new
trace file that combines the two networks.
 Our implementation will act as a bridge between the two networks and
the new trace file will show packets traveling from one network to the
other with transparency.
 The new trace file will not be a NS2 output, in other words, the satellitewireless network will still not work in NS2. But, this trace file will be the
closest simulation obtainable to a satellite-wireless network.
How are we going to reach our goal?
 The receiving terminal in the satellite network and the base station in the
wireless network will be clones of each other. That is, when the receiving
terminal receives a message from the satellite network, the base station
is going to route this message to the destination node in the wireless
network. On the same hand, when the base station receives a message
from the wireless network, the receiving terminal at the satellite network
will send this message to its destination node.
 The former paragraph suggests that the user specifies a source node in
the satellite network and a destination node in the wireless network, and
vice versa. So, we ask the user to add the following lines at the
beginning of each file:
# application
# “source” “destination” “start time” “interval/window” “packet size”
# “source” “destination” “start time” “interval/window” “packet size”
.
.
.
#.
# UDP
# node1 node1 3.0 50.0 500
# node2 node3 10.0 60.0 500
# TCP
# node4 node2 3.0 32 500
# node5 node5 10.0 32 500
#.
Let’s look at an example…
clones
node1
terminal1
terminal3
node2
node1
base_station1
terminal2
node3
node2
node3
node4
terminal4
node5
base_station2
terminal4
node6
WIRELESS NETWORK
SATELLITE NETWORK
node4
node5
Generating the final trace file
 The main problem that we will face in combining the trace files is that
both of them share same node names.
 To solve this problem we will change the node names in both trace files.
Satellite:
n01 … n999
Wireless:
n01 … n999
s01 … s999
w01 … w999
 We will merge both trace files and sort them according to time.
Everything will be executed from a script file, so the user will not worry about
running the source codes or our program to get the final trace file.