X. Creating Wired-cum-Wireless and MobileIP Simulations in ns

[Previous section] [Next section] [Back to the index]

IMPORTANT: This tutorial chapter uses new node APIs which are not available in the ns2.1b5 version. So please download the daily snapshot unless a release is made for version ns2.1b6 or higher. The current snapshot version is updated daily, so please check the validation test results for that day before downloading, as these snapshots can sometimes be unstable due to ongoing changes made by ns-developers.

X.1. Creating a simple wired-cum-wireless scenario

The wireless simulation described in section IX, supports multi-hop ad-hoc networks or wireless LANs. But we may need to simulate a topology of multiple LANs connected through wired nodes, or in other words we need to create a wired-cum-wireless topology.

In this section we are going to extend the simple wireless topology created in section IX to create a mixed scenario consisting of a wireless and a wired domain, where data is exchanged between the mobile and non-mobile nodes. We are going to make modifications to the tcl script called wireless1.tcl created in section IX.2 and name the resulting wired-cum-wireless scenario file wireless2.tcl.

For the mixed scenario, we are going to have 2 wired nodes, W(0) and W(1), connected to our wireless domain consisting of 3 mobilenodes (nodes 0, 1 & 2) via a base-station node, BS. Base station nodes are like gateways between wireless and wired domains and allow packets to be exchanged between the two types of nodes. For details on base-station node please see section 2 (wired-cum-wireless networking) of chapter 15 of ns notes&doc (now renamed as ns Manual). Fig1. shows the topology for this example described above.

Fig.1

Let us begin by checking what changes need to be made to the list of variables defined at the beginning of wireless1.tcl.

The Adhoc routing protocol is changed to DSDV. Also, we define TCP and CBR connections between the wired and wireless nodes in the script itself. So we won't need to use the connection pattern file used in earlier simulation. Also change the simulation stop time. Note here that we use array opt() instead of val() simply to illustrate that this is no longer a global array variable and its scope is defined only in the test script.
set opt(adhocRouting) DSDV set opt(cp) "" ;# cp file not used set opt(stop) 300 ;# time to stop simulation
We define the start times for TCP flows here:
set opt(ftp1-start) 160.0 set opt(ftp2-start) 170.0
Also add the following line to define number of wired and base-station nodes:
set num_wired_nodes 2 set num_bs_nodes 1

Now we move to the main part of the program. For mixed simulations we need to use hierarchical routing in order to route packets between wireless and wired domains. As explained in section 15.2.1 of ns Manual, in ns, the routing information for wired nodes are based on connectivity of the topology, i.e how are nodes connected to one another through Links. This connectivity information is used to populate the forwarding tables in each wired node. However wireless nodes have no concept of "links". Packets are routed in a wireless topology using their adhoc routing protocols which build forwarding tables by exchanging routing queries among its neighbours. So inorder to exchange pkts among these wired and wireless nodes, we use base-stations which act as gateways between the two domains. We seggregate wired and wireless nodes by placing them in different domains. Domains and sub-domains (or clusters as they are called here) are defined by means of hierarchical topology structure as shown below. After line "set ns [new Simulator]", add the following lines:
$ns_ node-config -addressType hierarchical AddrParams set domain_num_ 2 ;# number of domains lappend cluster_num 2 1 ;# number of clusters in each ;#domain AddrParams set cluster_num_ $cluster_num lappend eilastlevel 1 1 4 ;# number of nodes in each cluster AddrParams set nodes_num_ $eilastlevel ;# for each domain

In the above lines we first configure the node object to have addresstype as Hierarchical. Next the topology hierarchy is defined. Number of domains in this topology is 2 (one for the wired nodes and one for the wireless). Number of clusters in each of these domains is defined as "2 1" which indicates the first domain (wired) to have 2 clusters and the second (wireless) to have 1 cluster. The next line defines the number of nodes in each of these clusters which is "1 1 4"; i.e one node in each of the first 2 clusters (in wired domain) and 4 nodes in the cluster in the wireless domain. So the topology is defined into a 3-level hierarchy (see the topology figure above).

Next we setup tracing for the simulation. Note here that for wired-cum-wireless simulation traces may be generated for both wired and wireless domains. Both the traces are written into the same output file defined here as wireless2-out.tr. In order to differentiate wireless traces from wired ones, all wireless traces begin with "WL". We also setup nam traces. As mentioned earlier nam traces for wireless nodes currently show node movements only.
set tracefd [open wireless2-out.tr w] set namtrace [open wireless2-out.nam w] $ns_ trace-all $tracefd $ns_ namtrace-all-wireless $namtrace $opt(x) $opt(y)

Next we need to create the wired, wireless and base-station nodes. Note here that for all node creations, you have to pass the hierarchical address of the node. So after line "create-god $opt(nn)", add the following lines for creating wired nodes:
# create wired nodes set temp {0.0.0 0.1.0} ;# hierarchical addresses to be used for {set i 0} {$i < $num_wired_nodes} {incr i} { set W($i) [$ns_ node [lindex $temp $i]] }

In order to create base-station node, we need to configure the node structure as shown below. This is part of the new node API which consists of first configuring and then creating nodes. Refer to node API in chapterIX for details about the new node API. Since base-station nodes are gateways between wired and wireless domains they need to have wired routing mechanism turned on which is done by setting node-config option -wiredRouting ON. After creating the base-station node we reconfigure for wireless node and so turn wiredRouting OFF. All other node-config options used for base-station remains the same for mobilenode. Also the BS(0) node is assigned as the base-station node for all the mobilenodes in the wireless domain, so that all pkts originating from mobilenodes and destined outside the wireless domain, will be forwarded by mobilenodes towards their assigned base-station.

Note that it is important for the base-station node to be in the same domain as the wireless nodes. This is so that all pkts originating from the wired domain, and destined for a wireless node will reach the base-station which then uses its adhoc routing protocol to route the pkt to its correct destination. Thus in a mixed simulation involving wired and wireless nodes its necessary :
1) to turn on hierarchical routing
2) to create separate domains for wired and wireless nodes. There may be multiple wired and wireless domains to simulate multiple networks.
3) to have one base-station node in every wireless domain, thru which the wireless nodes may communicate with nodes outside their domain.

Let us go step by step for this example to see how the hierarchy is created. Here we have two domains, domain 0 , for wired and domain 1, for wireless. The two wired nodes are placed in 2 separate clusters, 0 and 1; thus their addresses look like 0(domain 0).0(cluster 0).0(only node) and 0 (same domain 0).1(cluster 1).0(again only node).
As for the wireless nodes, they are in domain 1; we have defined one cluster (0), so all nodes are in this cluster. Hence the addresses are:
Base-station: 1(second domain,1).0(cluster 0).0(first node in cluster)
WL node#1 : 1.0.1(second node in cluster)
WL node#2 : 1.0.2(third node)
WL node#3 : 1.0.3(fourth node)
We could have placed the two wired nodes in the same cluster in wired domain 0. Also we could have placed other wireless nodes in different clusters in wireless domain 1. Also depending on our topology we may have got rid of clusters altogether, and simply have had 2 layers of hierarchy, the domains and the nodes.

# configure for base-station node
$ns_ node-config -adhocRouting $opt(adhocRouting) \
                 -llType $opt(ll) \
                 -macType $opt(mac) \
                 -ifqType $opt(ifq) \
                 -ifqLen $opt(ifqlen) \
                 -antType $opt(ant) \
                 -propType $opt(prop) \
                 -phyType $opt(netif) \
                 -channelType $opt(chan) \
		 -topoInstance $topo \
                 -wiredRouting ON \
		 -agentTrace ON \
                 -routerTrace OFF \
                 -macTrace OFF 

#create base-station node
set temp {1.0.0 1.0.1 1.0.2 1.0.3}   ;# hier address to be used for
                                     ;# wireless domain
set BS(0) [ $ns_ node [lindex $temp 0]]
$BS(0) random-motion 0               ;# disable random motion

#provide some co-ordinates (fixed) to base station node
$BS(0) set X_ 1.0
$BS(0) set Y_ 2.0
$BS(0) set Z_ 0.0

# create mobilenodes in the same domain as BS(0)
# note the position and movement of mobilenodes is as defined
# in $opt(sc)
# Note there has been a change of the earlier AddrParams 
# function 'set-hieraddr' to 'addr2id'.

#configure for mobilenodes
$ns_ node-config -wiredRouting OFF

# now create mobilenodes
for {set j 0} {$j < $opt(nn)} {incr j} {
    set node_($j) [ $ns_ node [lindex $temp \
            [expr $j+1]] ]
    $node_($j) base-station [AddrParams addr2id \
            [$BS(0) node-addr]]   ;# provide each mobilenode with
                                  ;# hier address of its base-station
}

Next connect wired nodes and BS and setup TCP traffic between wireless node, node_(0) and wired node W(0), and between W(1) and node_(2), as shown below:
#create links between wired and BS nodes $ns_ duplex-link $W(0) $W(1) 5Mb 2ms DropTail $ns_ duplex-link $W(1) $BS(0) 5Mb 2ms DropTail $ns_ duplex-link-op $W(0) $W(1) orient down $ns_ duplex-link-op $W(1) $BS(0) orient left-down # setup TCP connections set tcp1 [new Agent/TCP] $tcp1 set class_ 2 set sink1 [new Agent/TCPSink] $ns_ attach-agent $node_(0) $tcp1 $ns_ attach-agent $W(0) $sink1 $ns_ connect $tcp1 $sink1 set ftp1 [new Application/FTP] $ftp1 attach-agent $tcp1 $ns_ at $opt(ftp1-start) "$ftp1 start" set tcp2 [new Agent/TCP] $tcp2 set class_ 2 set sink2 [new Agent/TCPSink] $ns_ attach-agent $W(1) $tcp2 $ns_ attach-agent $node_(2) $sink2 $ns_ connect $tcp2 $sink2 set ftp2 [new Application/FTP] $ftp2 attach-agent $tcp2 $ns_ at $opt(ftp2-start) "$ftp2 start"

This would be followed by the remaining lines from wireless1.tcl (sourcing cp and sc files, telling mobilenodes when to stop and finally running ns). It is possible that some lines of code in wireless2.tcl have not been discussed here. For a complete copy of script wireless2.tcl, download from here.

Run the script. The ns and nam trace files are generated at the end of simulation run. Running wireless2-out.nam shows the movement of mobilenodes and traffic in the wired domain. As mentioned earlier, traffic flow for mobilenodes is not as yet supported in nam. In trace file wireless2-out.tr we see traces for both wired domain and wireless domain (preceeding with "WL" for wireless). At 160.0s, a TCP connection is setup between _3_, (which is node_(0)) and 0, (which is W(0)). Note that the node-ids are created internally by the simulator and are assigned in the order of node creation. At 170s, another TCP connection is setup in the opposite direction, from the wired to the wireless domain. For details on CMUTraces see chapter 15 of ns documentation.

X.2. Running MobileIP in a simple wired-cum-wireless topology

So far we have created a wired-cum-wireless topology and have exchanged pkts between a wired and wireless domain via a base-station. But a mobilenode may roam outside the domain of its basestation and should still continue to receive packets destined to it. In other words it would be interesting to extend mobileIP support in the wired-cum-wireless scenario we created in section X.1.
For this example we have the same wired domain consisting of 2 wired nodes, W0 and W1. We have 2 base-station nodes and call them HomeAgent(HA) and ForeignAgent(FA) respectively. The wired node W1 is connected to HA and FA as shown in the figure below. There is a roaming mobilenode called MobileHost(MH) that moves between its home agent and foreign agents. We will set up a TCP flow between W0 and MH. As MH moves out from the domain of its HA, into the domain of FA, we will observe how pkts destined for MH is redirected by its HA to the FA as per mobileIP protocol definitions. See fig2 below for the topology described above.

Fig.2

We shall edit wireless2.tcl created in section X.1 to create the wireless-mip script called wireless3.tcl. It may be possible that the whole of wireless3.tcl is not discussed here. So for convinience, you may download a copy of wireless3.tcl from here.

Change number of mobilenodes and time of simulation,
set opt(nn) 1 ;# just one MH set opt(stop) 250

In this example we will set up the TCP connection as well as define movement of the MH in the script itself. Hence we are not going to use the cp and sc files.
set opt(cp) "" set opt(sc) ""

Define the TCP flow starttime,
set opt(ftp1-start) 100.0

Change number of wired, base-station and mobile nodes. However note that the variable num_bs_nodes is not really used in this script. The base-station nodes, HA and FA, are individually created and handled.
set num_wired_nodes 2

After the 2 lines creating ns instance and setting address format to hierarchical, add the following lines to define the topological hierarchy. It is quite similar to that of wireless2.tcl except that now we have a third domain for the FA. Change the cluster and node parameters accordingly.
AddrParams set domain_num_ 3 ;# number of domains lappend cluster_num 2 1 1 ;# number of clusters in each domain AddrParams set cluster_num_ $cluster_num lappend eilastlevel 1 1 2 1 ;# number of nodes in each cluster AddrParams set nodes_num_ $eilastlevel ;# of each domain

Next set up ns trace and nam files for wireless-mip,
set tracefd [open wireless3-out.tr w] set namtrace [open wireless3-out.nam w] $ns_ trace-all $tracefd $ns_ namtrace-all-wireless $namtrace $opt(x) $opt(y)

So in this topology we have one wired domain (denoted by 0) and 2 wireless domains (denoted by 1 & 2 respectively). Hence as described in section X.1, the wired node addresses remain the same, 0.0.0 and 0.1.0. In the first wireless domain (domain 1) we have base-station, HA and mobilenode, MH, in the same single cluster. Their addresses are 1.0.0 and 1.0.1 respectively. For the second wireless domain (domain 2) we have a base-station, FA with an address of 2.0.0. However in the course of the simulation, the MH will move into the domain of FA and we shall see how pkts originating from a wired domain and destined to MH will reach it as a result of the MobileIP protocol.

Wired nodes will be created as earlier. However in place of a single base-station node, a HA and FA will be created. Note here that to turn the mobileIP flag on we have configure the node structure accordingly using option -mobileIP ON.
# Configure for ForeignAgent and HomeAgent nodes $ns_ node-config -mobileIP ON \ -adhocRouting $opt(adhocRouting) \ -llType $opt(ll) \ -macType $opt(mac) \ -ifqType $opt(ifq) \ -ifqLen $opt(ifqlen) \ -antType $opt(ant) \ -propType $opt(prop) \ -phyType $opt(netif) \ -channelType $opt(chan) \ -topoInstance $topo \ -wiredRouting ON \ -agentTrace ON \ -routerTrace OFF \ -macTrace OFF # Create HA and FA set HA [$ns_ node 1.0.0] set FA [$ns_ node 2.0.0] $HA random-motion 0 $FA random-motion 0 #provide some co-ord (fixed) to these base-station nodes. $HA set X_ 1.000000000000 $HA set Y_ 2.000000000000 $HA set Z_ 0.000000000000 $FA set X_ 650.000000000000 $FA set Y_ 600.000000000000 $FA set Z_ 0.000000000000

Next create the mobilehost as follows. Note as before we have to turn off the option -wiredRouting (used for creation of base-station nodes) before creating mobilenodes. Also the HA is setup as the home-agent for the mobilehost. The MH has an address called the care-of-address (COA). Based on the registration/beacons exchanged between the MH and the base-station node (of the domain the MH is currently in), the base-station's address is assigned as the MH's COA. Thus in this simulation, address of HA is assigned initially as the COA of MH. As MH moves in to the domain of FA, its COA changes to that of the FA. For details on MobileIP implementation in ns, read section 15.2.2 of (in wireless networking chapter) ns documentation. Also see files mip.{cc,h}, mip-reg.{cc,h}, tcl/lib/{ns-mip.tcl, ns-wireless-mip.tcl}.
# configure for mobilehost $ns_ node-config -wiredRouting OFF # create mobilehost that would be moving between HA and FA. # note address of MH indicates its in the same domain as HA. set MH [$ns_ node 1.0.1] set node_(0) $MH set HAaddress [AddrParams set-hieraddr [$HA node-addr]] [$MH set regagent_] set home_agent_ $HAaddress # movement of the MH $MH set Z_ 0.000000000000 $MH set Y_ 2.000000000000 $MH set X_ 2.000000000000 # MH starts to move towards FA (640, 610) at a speed of 20m/s $ns_ at 100.000000000000 "$MH setdest 640.000000000000 610.000000000000 20.000000000000" # and goes back to HA (2, 2) at a speed of 20 m/s $ns_ at 200.000000000000 "$MH setdest 2.000000000000 2.000000000000 20.000000000000"

Create links between Wired nodes and HA/FA and setup TCP connection:
# create links between wired and BaseStation nodes $ns_ duplex-link $W(0) $W(1) 5Mb 2ms DropTail $ns_ duplex-link $W(1) $HA 5Mb 2ms DropTail $ns_ duplex-link $W(1) $FA 5Mb 2ms DropTail $ns_ duplex-link-op $W(0) $W(1) orient down $ns_ duplex-link-op $W(1) $HA orient left-down $ns_ duplex-link-op $W(1) $FA orient right-down # setup TCP connections between a wired node and the MobileHost set tcp1 [new Agent/TCP] $tcp1 set class_ 2 set sink1 [new Agent/TCPSink] $ns_ attach-agent $W(0) $tcp1 $ns_ attach-agent $MH $sink1 $ns_ connect $tcp1 $sink1 set ftp1 [new Application/FTP] $ftp1 attach-agent $tcp1 $ns_ at $opt(ftp1-start) "$ftp1 start"

The rest of the script remains unchanged (i.e tell mobilenodes when the simulation stops). Save and run the script. Click here for a copy of the file wireless3.tcl.

While running the script, you may see warnings like "warning: Route to base_stn not known: dropping pkt". This means that as the MH moves from the domain of one base-station into domain of another there may be interim periods when it is not registered to any base-station and thus doesnot know whom to forward pkts destined outside its domain. On completion of the run, ns and nam trace output files "wireless3-out.tr" and "wireless3-out.nam" are created. The nam output shows the movement of the mobilehost and traffic flow in the wired domain. The ns trace output shows traces for both the wired nodes as well as the wireless domain. We see routine beacon broadcast/solicitations sent out by HA/FA and the MH. Initially the TCP pkts are handed down to MH directly by its HA. As MH moves away from HA domain into the domain of the FA, we find the pkts destined for MH, being encapsulated and forwarded to the FA which then strips off or decapsulates the pkt and hands it over to the MH.

[Previous section] [Next section] [Back to the index]

VINT

[email protected]