NS2模拟Ping

http://www.isi.edu/nsnam/ns/tutorial/跟着这个添加自己的ping协议，环境是Ubuntu10.04, NS2.34,其实这个版本的NS已经实现的ping协议，并且与前面网站版本不同，在改代码的时候会有所不同。首先添加自己的ping.h头文件：

 1 #ifndef ns_ping_h
 2 #define ns_ping_h
 3 
 4 #include "agent.h"
 5 #include "tclcl.h"
 6 #include "packet.h"
 7 #include "address.h"
 8 #include "ip.h"
 9 
10 struct hdr_ping {
11     //set '0' if the packet is on its way from the sender to the node which is being pinged,while it is going to
12 //be set to '1'on its way back.
13     char ret;
14     //a time stamp that is set on the packet when it is sent.
15     double send_time;
16 }
17 
18 class PingAgent : public Agent {
19     public :
20         PingAgent();
21         int command(int argc, const char * const * argv);
22         void recv(Packet *, Handler *);
23     protected :
24         int off_ping_;
25 }
26 #endif

注意为了不影响已经存在的分组头，最好是自己添加命名规则一般是hdr_new。

 1 static class PingHeaderClass : public PacketHeaderClass {
 2     public :
 3         PingHeaderClass() : PacketHeaderClass("PacketHeader/Ping",
 4                     sizeof(hdr_ping)) {}
 5 }class_pinghdr;
 6 
 7 static class PingClass : public TclClass {
 8     public :
 9         PingClass() : TclClass("Agent/Ping") {}
10         TclObject *create(int, const char * const *) {
11             return (new PingAgent());
12         }
13 }class_ping;
14 
15 PingAgent :: PingAgent() : Agent(PT_PING)
16 {
17     bind("packetSize_", &size);
18     bind("off_ping_", &off_ping_);
19 }
20 
21 int PingAgent :: command(int argc, const *char *const argv)
22 {
23     if(argc == 2) {
24         if(strcmp(argv[1], "send") == 0) {
25             //create a new packet
26             Packet* pkt = allocpkt();
27             //access the ping header for tne new packet
28             hdr_ping* hdr = (hdr_ping*)pkt->access(off_ping_);
29             //set the 'ret'    field to 0, so the receiving node knows
30 //that it has to generate an echo packet
31             hdr->ret = 0;
32             //store the current time in the 'send_time' field
33             hdr->send_time = Schedulet::instance.clock();
34             //send the packet
35             send(pkt, 0);
36             //return TCL_OK, so the calling function knows that 
37 //the command has been processed
38             return TCL_OK;
39         }
40     }
41 
42     //if the command hasn't been processed by PingAgent()::command,
43 //call the command() function for the base class
44     return (Agent::command(argc, argv));
45 }
46 
47 
48 void PingAgent :: recv(Pakcet *pkt, Handler *)
49 {
50     //aceess the IP header for the recived packet:
51     hdr_ip * hdrip = (hdr_ip*)pkt->access(off_ip_);
52     //access the ping header for the recived packet:
53     hdr_ping * hdr = (hdr_ping*)pkt->access(off_ping_);
54 
55     if(hdr->ret == 0) {
56         //send an 'echo'.First save the old packet's send_time/
57         double stime = hdr->send_time;
58         //discard the packet
59         Packet::free(pkt);
60         //create a new packet
61         Packet * pktret = allocpkt();
62         //access the ping header for the new packet
63         hdr_ping* hdrret = (hdr_ping*)pktret->access(off_ping_);
64         //set the 'ret' field to 1, so the receiver won't send another echo 
65         hdrret->ret = 1;
66         //set the send time field to the correct value
67         hdrret->send_time = stime;
68         //send the packet
69         send(pktret, 0);
70     } else {
71         //a packet was received. Use tcl.eval to call the Tcl
72 //interreter with the ping results.
73 //note : In the Tcl code, a procedure 'Agent/Ping recv {form rtt}'
74 //has to be defined which allows the user to react to the ping
75 //result.
76         char out[100];
77         //prepare the output to the Tcl interpreter. Calculate the round
78 //trip time.
79         sprintf(out, "%s recv %d %3.1f", name(), 
80                     hdrip->src_.addr_ >> Address::instance().NodeShitf_[1],
81                     (Scheduler::instance().clock()-hdr->send_time) * 1000);
82         Tcl& tcl = Tcl::instance();
83         tcl.eval(out);
84         //discard the packet
85         Packet::free(pkt);
86     }
87 }

有4个C++类与分组头和分组有关，它们分别是Packet, p_info, PacketHeaderClass和PacketHeaderManager.其中 PacketHeaderClass是各种分组头的基类，而它本身是继承于TclClass的，现在对PingHeaderClass的作用理解不是很清楚，而PingClass用于创建一个PingAgent对象，注意关键字static，这样保证了当NS2启动时能调用create返回一个TclObject对象（NS2的分裂对象模型），PingAgent的构造函数绑定了两个变量。command方法牵涉到cmd机制，cmd命令是meet the Tcl Unknown mechanism——Tcl的unknown机制，简单来说，当向ns输入一个未知的命令时，Tcl的unknow机制就会调用cmd函数，cmd函数激活shadow对象的command方法，进而执行这个command方法，在这里我们在command中添加了send命令。在这里我们并没有重写父类Agent的send方法，可以在~ns/common/agent.{cc, h}中看到send的实现：

1 void send(Packet* p, Handler* h) { target_->recv(p, h); }

所以，send方法调用了接受者的recv方法，而我们实现了recv方法，并调用tcl对象的recv方法，这就是为什么在下面的tcl文件中recv方法被调用的理由。

当然，还需要一些必要的改变：

首先要增加分组头类型，在NS2.34中，修改了packet_t的类型，在ns2.33之前packet_t是enum，而之后定义为unsigned int.所以添加如下：

1 static const packet_t PT_PING = 44;

需要注意的是，ns2.34已经实现了ping协议，所以我并没有添加任何内容，如果是自己另写协议，在添加packet_t时，必须注释掉不用的分组头，并且要按所有常量值的顺序存储，不然就会有如下错误：

1 diffserv -I./satellite -I./wpan -I./mit/rca -I./mit/uAMPS -o common/ptypes2tcl.o common/ptypes2tcl.cc
2 g++ -Wl,-export-dynamic  -o common/ptypes2tcl common/ptypes2tcl.o
3 ./common/ptypes2tcl > gen/ptypes.cc
4 Segmentation fault
5 make: *** [gen/ptypes.cc] Error 139

PS：还需要该/bin目录的权限：

chmod -R 755 /bin

不然会有权限问题，还得用sudo  make。

在p_info类中添加内容时也可网站不同：

1 p_info()
2      {
3          initName();
4      }

所以需要在initName()方法中添加名字。另外，需要在~ns/tcl/lib/ns-default.tcl中添加包的大小：

1 Agent/Ping set packetSize_ 64

然后在~ns/tcl/lib/ns-packet.tcl的foreach prot模块中添加：

1 Ping    # Ping

最后在makefile文件中添加，直接复制源码吧（自己以后注意添加就是了）：

1 tcp/tfrc.o tcp/tfrc-sink.o mobile/energy-model.o apps/ping.o tcp/tcp-rfc793edu.o

可以看到，ns自带的ping源码在apps/目录下。

再加上tcl文件：

 1 #create a simulator object
 2 set ns [new Simulator]
 3 
 4 #open a trace file
 5 set tracefd [open out.tr w]
 6 $ns trace-all $tracefd
 7 set nf [open out.nam w]
 8 $ns namtrace-all $nf
 9 
10 #define a finish procedure
11 proc finish {} {
12     global ns nf
13     $ns flush-trace
14     close $nf
15     exec nam out.nam &
16     exit 0
17 }
18 
19 #create three nodes
20 set n0 [$ns node]
21 set n1 [$ns node]
22 set n2 [$ns node]
23 
24 #connect the nodes with two links
25 $ns duplex-link $n0 $n1 1Mb 10ms DropTail
26 $ns duplex-link $n1 $n2 1Mb 10ms DropTail
27 
28 #define a recv funciton for the class Agent/Ping
29 Agent/Ping instproc recv {from rtt} {
30     $self instvar node_
31     puts "node [$node_ id] received ping answer from/
32         $from with round-trip-time $rtt ms."
33     }
34 
35 #create two ping agents and attach them to the nodes n0 and n2
36 set p0 [new Agent/Ping]
37 $ns attach-agent $n0 $p0
38 
39 set p1 [new Agent/Ping]
40 $ns attach-agent $n2 $p1
41 
42 #connect
43 $ns connect $p0 $p1
44 
45 $ns at 0.2 "$p0 send"
46 $ns at 0.4 "$p1 send"
47 $ns at 0.6 "$p0 send"
48 $ns at 0.6 "$p1 send"
49 $ns at 1.0 "finish"
50 
51 $ns run

可以看到在tcl文件中，调用了send命令，并且添加了tcl对象的recv方法，方便c++shadow对象调用。需要注意的是，connect方法并没有建立连接，只是告知目的地址，并且在ns中，应用层之间并没有发送数据，都是由代理模拟的。

疑问：对NS2的内部运行机制还是半懂不懂，并且不理解单线程是怎么实现多节点同时发送数据的，另外，对PacketHeaderManager类的作用也还没有体会到。