1. 使用Ping做什么
ping用於確定本地主機是否能與另一台主機成功交換(發送與接收)數據包,再根據返回的信息,就可以推斷TCP/IP參數是否設置正確,以及運行是否正常、網絡是否通暢等。
2. 效果
CMD命令:
Python程序:
3. 在驗證兩台主機是否能正常聯通時做了什么
驗證的過程就像打電話一樣,我們如果要知道自己能否給另一個人交流就給他打一個電話,如果你聽到了對方的聲音那就是打通了,否則就是失敗。但是無論成功與否我們都有一個這樣的過程:
通過電話線(或是無線,即手機)給對方發送一個請求通話的消息,無論對方是否接通都會返回一條應答的消息。
這里我們發送或接收的消息就是報文(IP報文),其中,發送的請求通話消息我們稱為“請求回顯”(Ping請求),而接收到的應答消息稱為“回顯應答”(Ping應答)。
我們通過“回顯應答”類型報文中的內容來判斷是否能夠正常連接(交換數據包)。
而在Ping過程中需要進行四次(並不固定,一般默認四次)請求,這個可以看做是連續打四次電話。
4. IP報文中有什么
由於IP協議並不是一個可靠的協議,它不保證數據被成功送達,那么,如何才能保證數據的可靠送達呢? 這里就需要使用到一個重要的協議模塊ICMP(網絡控制報文)協議。它傳遞差錯報文以及其他需要注意的信息,經常供IP層或更高層協議(TCP或UDP)使用。所以它經常被認為是IP層的一個組成部分。它在IP數據報文中的封裝如下:
在這里我們不關心IP首部內容(只需要記住它有20bytes即可),而ICMP報文的類型正如上述所說有好多種,我們根據報文的type和code來判斷ICMP報文的類型(當然這里我們也只關系回顯請求和回顯應答):
ICMP報文 = ICMP報頭+Data(這個消息可以自定義,當然每個不同類型的報文的data都會有所不同)
(ICMP_Packet = ICMPHeader + Data)
5. ICMP報文中的各個數據是什么,有什么用
(1)type和code
type和code決定了ICMP報文是什么類型(當然這里我們也只關系回顯請求和回顯應答)。
以下是不同type(ICMP類型)和code(編碼)對應的不同ICMP報文類型,不用全部記住,在這里我們只要知道:
type = 8, code = 0 表示 回顯請求
type = 0, code = 0 表示 回顯應答
(2)Checksum(校驗和)
用於檢驗數據包是否正確。
(3)ID(標識符)
每一個不同的進程都有一個ID作為標識,即每次運行這個Ping程序都會獲得一個進程ID,我們通過這個比較發送進程ID和回顯應答中的ID,如果一樣則說明是同一個Ping程序
(4)sequence(序列號)
每發送一次消息,就有一個序列號,第一個就是1,第二次就是2等等(也可以從“0”開始,可以自己定義),按照收發四次來講就是0、1、2、3
(5)Data(數據)
你可以將你自己想要的其他信息放到這里,比如發送消息的時間,這樣就可以計算延遲了
6. 程序框架及具體代碼
首先,我們知道要做到兩台host通信需要借助socket,以下內容也是在此基礎上展開的
(1)全局變量
設置了幾個全局變量,包括:回顯請求(8),回顯應答(0),Ping標識符(ID),Ping序列號(sequence)
ICMP_ECHO_REQUEST = 8 # ICMP type code for echo request messages ICMP_ECHO_REPLY = 0 # ICMP type code for echo reply messages ID = 0 # ID of icmp_header SEQUENCE = 0 # sequence of ping_request_msg
(2)main()
首先我們要有這個程序的入口,俗稱“主函數”
在這里我們調用Ping(host, timeout = 1)【自定義】函數正式開始這個過程
ping("baidu.com")
(3)Ping(host, timeout = 1)
有兩個參數(host,timeout):
host是要嘗試連接的主機的域名或者直接寫IP地址也沒問題
timeout是設置的超時時間,當等待時間超過1s時我們認為連接超時,即連接失敗(表示無法正常交流)
內容簡介:
定義變量---在Ping函數中我們設置了幾個變量用於表示發送、成功接收、丟失包的數量 和 最長時延、最短時延、總時延
調用doOnePing【自定義】函數---這個函數包括對發送和接收報文消息的函數,返回值就是時延
此外---當然,我們對doOnePing函數調用了四次
def ping(host, timeout=1): send = 0 lost = 0 receive = 0 maxTime = 0 minTime = 1000 sumTime = 0 # 1. Look up hostname, resolving it to an IP address desIp = socket.gethostbyname(host) global ID ID = os.getpid() for i in range(0, 4): global SEQUENCE SEQUENCE = i # 2. Call doOnePing function, approximately every second delay = doOnePing(desIp, timeout) * 1000 send += 1 if delay > 0: receive += 1 if maxTime < delay: maxTime = delay if minTime > delay: minTime = delay sumTime += delay # 3. Print out the returned delay print("Receive from: " + str(desIp) + ", delay = " + str(int(delay)) + "ms") else: lost += 1 print("Fail to connect.") time.sleep(1) # 4. Continue this process until stopped avgTime = sumTime / receive recvRate = receive / send * 100.0 print("\nSend: " + str(send) + ", success: " + str(receive) + ", lost: " + str(lost) + ", rate of success: " + str(recvRate) + "%.") print( "MaxTime = " + str(int(maxTime)) + "ms, MinTime = " + str(int(minTime)) + "ms, AvgTime = " + str(int(avgTime)))
(4) doOnePing(destination, timeout)
有兩個參數(destination,timeout):
destination就是獲得的目標主機的IP地址
timeout就是超時時間(自定義的)
def doOnePing(destinationAddress, timeout): # 1. Create ICMP socket icmpName = socket.getprotobyname('icmp') icmp_Socket = socket.socket(socket.AF_INET, socket.SOCK_RAW, icmpName) # 2. Call sendOnePing function sendOnePing(icmp_Socket, destinationAddress, ID) # 3. Call receiveOnePing function totalDelay = receiveOnePing(icmp_Socket, destinationAddress, ID, timeout) # 4. Close ICMP socket icmp_Socket.close() # 5. Return total network delay return totalDelay
內容簡介:
創建socket---正如上述所說,我們通過socket來做這個通信,由於是使用ICMP協議,我們需要提前獲取這個協議(根據協議名)
調用sendOnePing---發送一條回顯請求給目標主機
調用receiveOnePing---接收回顯應答函數,返回值為時延
關閉socket
返回在調用receiveOnePing函數后獲得的時延
(5)sendOnePing(icmpSocket, destination, ID)
參數有三個,用於做什么很顯然了,這里不再贅述
def sendOnePing(icmpSocket, destinationAddress, ID): icmp_checksum = 0 # 1. Build ICMP header icmp_header = struct.pack('!bbHHh', ICMP_ECHO_REQUEST, 0, icmp_checksum, ID, SEQUENCE) time_send = struct.pack('!d', time.time()) # 2. Checksum ICMP packet using given function icmp_checksum = checksum(icmp_header + time_send) # 3. Insert checksum into packet icmp_header = struct.pack('!bbHHh', ICMP_ECHO_REQUEST, 0, icmp_checksum, ID, SEQUENCE) # 4. Send packet using socket icmp_packet = icmp_header + time_send icmpSocket.sendto(icmp_packet, (destinationAddress, 80)) # 5. Record time of sending
內容簡介:
這個函數的目的就在於向目標主機發送一條回顯請求報文,而這個報文中應該包括:ICMP_Header+Data(這里是發送的時間,用於計算時延)
其中用到的struct.pack函數請自行百度,如果不想知道細節的話只需要記住pack的第一個參數規定了其編碼格式(有各種字母符號的映射,可以自己編排),而之后的參數都是根據這個編碼格式進行打包的數據
在這個函數中調用了校驗和程序,這個原理也自行百度(很容易找到),不做深究的話也可以直接拿來用
函數編寫步驟:
* 將ICMP_Header的那幾個打包成為數據包(其中的校驗和只是為了Header的完整性而隨意設置的,占位用)
* 用校驗和程序檢測數據包的完整性
* 加入真正的校驗和(這里“真正”似乎不太正確,但暫時想不出有什么其他的詞匯形容了)重新打包
* 將發送的時間(time.time())編碼打包,作為ICMP的Data
* ICMP_Packet = ICMP-Header + Data
* 借助socket發送報文
* 此外,sendto(msg, (IP, Port))中的Port為80是因為Web默認使用端口號80(還有許多網絡應用有特定的端口號——方便訪問,一般靠前的port都是分配好的,如果是自己構建的程序則可以申請使用空閑的port)
(6)receiveOnePing(icmpSocket, destination, ID, timeout)
這里參數不再介紹
def receiveOnePing(icmpSocket, destinationAddress, ID, timeout): # 1. Wait for the socket to receive a reply timeBeginReceive = time.time() whatReady = select.select([icmpSocket], [], [], timeout) timeInRecev = time.time() - timeBeginReceive if not whatReady[0]: print("none") return -1 # 2. Once received, record time of receipt, otherwise, handle a timeout recPacket, addr = icmpSocket.recvfrom(1024) timeReceived = time.time() # 3. Compare the time of receipt to time of sending, producing the total network delay byte_in_double = struct.calcsize("!d") timeSent = struct.unpack("!d", recPacket[28: 28 + byte_in_double])[0] totalDelay = timeReceived - timeSent # 4. Unpack the packet header for useful information, including the ID rec_header = recPacket[20:28] replyType, replyCode, replyCkecksum, replyId, replySequence = struct.unpack('!bbHHh', rec_header) # 5. Check that the ID matches between the request and reply if ID == replyId and replyType == ICMP_ECHO_REPLY: # 6. Return total network delay return totalDelay if timeInRecev > timeout: print('overtime') return -1
內容簡介:
在這里我們統一規定如果成功接收了就返回delay(時延),如果沒有就返回-1(也可以是其他的自定義數值,只要能區分就可以)
select.select函數自行了解:whatReady[0] 用於判斷是否接收到回顯應答,返回true/false,應答不為空就是成功(繼續計算時延),否則就是失敗(直接返回-1跳出)
其中的timeInRecev用於判斷超時
我們在一開始就說了,IPHeader為20bytes,所以ICMPHeader應該從20開始截取(str[20:28]為ICMP_Header部分)
獲取發送時間的那一部分中涉及到了struct.calcsize()函數,可以看這個鏈接中的內容做基本了解:https://blog.51cto.com/firefish/112690:因為unpack函數解碼的內容至少要是8bytes,這個函數可以考慮為占位用的
7. 完整代碼
#!/usr/bin/python # -*- coding: UTF-8 -*- import os import struct import time import select import socket ICMP_ECHO_REQUEST = 8 # ICMP type code for echo request messages ICMP_ECHO_REPLY = 0 # ICMP type code for echo reply messages ICMP_Type_Unreachable = 11 # unacceptable host ICMP_Type_Overtime = 3 # request overtime ID = 0 # ID of icmp_header SEQUENCE = 0 # sequence of ping_request_msg def checksum(strings): csum = 0 countTo = (len(strings) / 2) * 2 count = 0 while count < countTo: thisVal = strings[count + 1] * 256 + strings[count] csum = csum + thisVal csum = csum & 0xffffffff count = count + 2 if countTo < len(strings): csum = csum + strings[len(strings) - 1] csum = csum & 0xffffffff csum = (csum >> 16) + (csum & 0xffff) csum = csum + (csum >> 16) answer = ~csum answer = answer & 0xffff answer = answer >> 8 | (answer << 8 & 0xff00) return answer def receiveOnePing(icmpSocket, ID, timeout): # 1. Wait for the socket to receive a reply timeBeginReceive = time.time() whatReady = select.select([icmpSocket], [], [], timeout) timeInRecev = time.time() - timeBeginReceive if not whatReady[0]: return -1 timeReceived = time.time() # 2. Once received, record time of receipt, otherwise, handle a timeout recPacket, addr = icmpSocket.recvfrom(1024) # 3. Compare the time of receipt to time of sending, producing the total network delay byte_in_double = struct.calcsize("!d") timeSent = struct.unpack("!d", recPacket[28: 28 + byte_in_double])[0] totalDelay = timeReceived - timeSent # 4. Unpack the packet header for useful information, including the ID rec_header = recPacket[20:28] replyType, replyCode, replyCkecksum, replyId, replySequence = struct.unpack('!bbHHh', rec_header) # 5. Check that the ID matches between the request and reply if ID == replyId and replyType == ICMP_ECHO_REPLY: # 6. Return total network delay return totalDelay elif timeInRecev > timeout or replyType == ICMP_Type_Overtime: return -3 # ttl overtime/timeout elif replyType == ICMP_Type_Unreachable: return -11 # unreachable else: print("request over time") return -1 def sendOnePing(icmpSocket, destinationAddress, ID): icmp_checksum = 0 # 1. Build ICMP header icmp_header = struct.pack('!bbHHh', ICMP_ECHO_REQUEST, 0, icmp_checksum, ID, SEQUENCE) time_send = struct.pack('!d', time.time()) # 2. Checksum ICMP packet using given function icmp_checksum = checksum(icmp_header + time_send) # 3. Insert checksum into packet icmp_header = struct.pack('!bbHHh', ICMP_ECHO_REQUEST, 0, icmp_checksum, ID, SEQUENCE) # 4. Send packet using socket icmp_packet = icmp_header + time_send icmpSocket.sendto(icmp_packet, (destinationAddress, 80)) def doOnePing(destinationAddress, timeout): # 1. Create ICMP socket icmpName = socket.getprotobyname('icmp') icmp_Socket = socket.socket(socket.AF_INET, socket.SOCK_RAW, icmpName) # 2. Call sendOnePing function sendOnePing(icmp_Socket, destinationAddress, ID) # 3. Call receiveOnePing function totalDelay = receiveOnePing(icmp_Socket, ID, timeout) # 4. Close ICMP socket icmp_Socket.close() # 5. Return total network delay return totalDelay def ping(host, count, timeout): send = 0 lost = 0 receive = 0 maxTime = 0 minTime = 1000 sumTime = 0 # 1. Look up hostname, resolving it to an IP address desIp = socket.gethostbyname(host) global ID ID = os.getpid() for i in range(0, count): global SEQUENCE SEQUENCE = i # 2. Call doOnePing function, approximately every second delay = doOnePing(desIp, timeout) * 1000 send += 1 if delay > 0: receive += 1 if maxTime < delay: maxTime = delay if minTime > delay: minTime = delay sumTime += delay # 3. Print out the returned delay print("Receive from: " + str(desIp) + ", delay = " + str(int(delay)) + "ms") else: lost += 1 print("Fail to connect. ", end="") if delay == -11: # type = 11, target unreachable print("Target net/host/port/protocol is unreachable.") elif delay == -3: # type = 3, ttl overtime print("Request overtime.") else: # otherwise, overtime print("Request overtime.") time.sleep(1) # 4. Continue this process until stopped if receive != 0: avgTime = sumTime / receive recvRate = receive / send * 100.0 print( "\nSend: {0}, success: {1}, lost: {2}, rate of success: {3}%.".format(send, receive, lost, recvRate)) print( "MaxTime = {0}ms, MinTime = {1}ms, AvgTime = {2}ms".format(int(maxTime), int(minTime), int(avgTime))) else: print("\nSend: {0}, success: {1}, lost: {2}, rate of success: 0.0%".format(send, receive, lost)) if __name__ == '__main__': while True: try: hostName = input("Input ip/name of the host you want: ") count = int(input("How many times you want to detect: ")) timeout = int(input("Input timeout: ")) ping(hostName, count, timeout) break except Exception as e: print(e) continue
8. 寫在最后
寫的時候翻了不少文章,但是都找不到了,只有這個離寫這篇文章的時間最近,就貼上來
參考:https://www.cnblogs.com/JetpropelledSnake/p/9177770.html