1、Chapter 6 網(wǎng)絡(luò)技術(shù),6.1 Data Link Layer 6.2 Network Layer 6.3 Network and Internetwork Devices 6.4 Virtual-Circuit Networks 6.5 Summary,6.1 數(shù)據(jù)鏈路層,6.1.1 數(shù)據(jù)鏈路層的功能,為網(wǎng)絡(luò)層提供服務(wù) 成幀 差錯(cuò)控制 流量控制,1.為網(wǎng)絡(luò)層提供服務(wù),基本服務(wù):將源機(jī)器中來自網(wǎng)絡(luò)層的數(shù)據(jù)傳輸給目的機(jī)器的網(wǎng)絡(luò)層 . (figure 6.1),2. 成幀,字符計(jì)數(shù)法 帶字符填充的首尾界符法 帶位填充的首位標(biāo)志法 物理層編碼違例法,種成幀方法：,在幀頭部使用一個(gè)字段標(biāo)明幀內(nèi)字

2、符數(shù)(figure 6.3),用DLE字符填充(figure 6.4),字符計(jì)數(shù)法,帶字符填充的首位界符法,每一幀以特殊的字符開始和結(jié)束 。如STX 、ETX,問題: 計(jì)數(shù)會(huì)被傳輸錯(cuò)誤完全打亂 (figure 6.3),如果這些特殊字符在數(shù)據(jù)中出現(xiàn)怎么辦？,用DLE字符轉(zhuǎn)義,開始和結(jié)束標(biāo)志字節(jié) : 01111110. 發(fā)送方遇到數(shù)據(jù)中有五個(gè)連續(xù)“1”時(shí)，自動(dòng)在其后插入“0” . 接收方遇到數(shù)據(jù)中有五個(gè)連續(xù)“1”時(shí)并且后面跟著一個(gè)0，自動(dòng)將“0” 取出丟棄. (figure 6.5),帶位填充的首尾標(biāo)志法,物理層編碼違例法,只適用于在物理介質(zhì)的編碼策略中采用冗余技術(shù)的網(wǎng)絡(luò),3. 差錯(cuò)控制,4.

3、流量控制,返回確認(rèn)幀以通告接收情況 設(shè)置計(jì)時(shí)器來避免永久等待 對(duì)幀編號(hào)避免重復(fù)幀,基于反饋機(jī)制的流量控制 基于速率的流量控制,6.1.2 基本數(shù)據(jù)鏈路協(xié)議,Appendix A,1. 無限制的單工協(xié)議,假設(shè):,數(shù)據(jù)只做單向傳輸 雙方網(wǎng)絡(luò)層總處于就緒狀態(tài) 緩沖空間無限大 交互信道為理想信道（無差錯(cuò)）,發(fā)送方每發(fā)出一幀后就等待一個(gè)應(yīng)答幀。 只有在接收到應(yīng)答信號(hào)后，才發(fā)送下一幀figure 6.6,Appendix B,2.停-等協(xié)議,優(yōu)點(diǎn): 簡單 缺點(diǎn): 效率低,3.停等自動(dòng)重復(fù)請求協(xié)議,發(fā)送設(shè)備在接收到最近幀的應(yīng)答消息前必須保留該幀的備份,為識(shí)別各幀，數(shù)據(jù)幀和應(yīng)答幀(ACK)都必須交替地標(biāo)識(shí)為0

4、 和1,否定應(yīng)答幀(NAK)是不編號(hào)的，它通知發(fā)送方重新發(fā)送 最近的一幀。,發(fā)送器安裝有一個(gè)定時(shí)器,損壞幀 接收方返回一個(gè)否定應(yīng)答幀(NAK)給發(fā)送方，發(fā)送方就重發(fā)該數(shù)據(jù)幀 。 (figure 6.7),丟失幀 信息幀丟失 發(fā)送方有一個(gè)定時(shí)器，每次發(fā)送一個(gè)數(shù)據(jù)幀就啟動(dòng)它。(figure 6.8),丟失確認(rèn)幀 發(fā)送設(shè)備一直等到定時(shí)器超時(shí)，然后重新發(fā)送數(shù)據(jù)幀。 (figure 6.9),Appendix C :A retransmission protocol with positive acknowledgement ACK,幀錯(cuò)誤的處理,6.1.3 滑動(dòng)窗口,1. 滑動(dòng)窗口 窗口:指一個(gè)發(fā)送方

5、和接收方都要?jiǎng)?chuàng)建的額外緩沖區(qū),識(shí)別機(jī)制 : 幀以模n方式標(biāo)號(hào)，也就是說從0到n-1編號(hào) For example 如果n=8，幀標(biāo)號(hào)就為：0，1，2，3，4，5，6，7,在傳輸?shù)拈_始，發(fā)送方窗口有n-l幀 . 隨著數(shù)據(jù)幀的傳送出去，當(dāng)前指針 S向右移動(dòng)。 一旦一個(gè)應(yīng)答幀(ACK)到來，窗口根據(jù)應(yīng)答幀(ACK)中應(yīng)答幀的個(gè)數(shù)對(duì)窗口進(jìn)行相同數(shù)目的擴(kuò)展 . 窗口大小不大于 n-1. (figure 6.11 ),發(fā)送窗口(figure 6.10),接收窗口大小為 1. 當(dāng)數(shù)據(jù)幀達(dá)到接收端，窗口向右移動(dòng)。 接收端按照特定的順序接收數(shù)據(jù)。 R 是期望的下一幀序號(hào).,一個(gè)3幀窗口的滑動(dòng)窗口(figure 6

6、.13),接收窗口 (figure 6.12),滑動(dòng)窗口自動(dòng)重傳請求協(xié)議,在發(fā)送出去的幀被應(yīng)答消息確認(rèn)前保持它們所有的備份,ACK幀帶有期待的幀的編號(hào)，NAK幀則帶有損壞幀的編號(hào) .,定時(shí)器,2. 回退N幀協(xié)議,幀損壞 NAK 有兩個(gè)含義: 在被破壞的幀之前的所有幀都正確接收了 . 所指示的幀的否定應(yīng)答 (figure 6.14),數(shù)據(jù)幀丟失 發(fā)送設(shè)備就重新傳輸由否定應(yīng)答幀(NAK)指明的幀，以及該丟失幀后已經(jīng)發(fā)送的所有幀。 (figure 6.15),確認(rèn)幀丟失 采用定時(shí)器 (figure 6.16),特點(diǎn)：如果一幀丟失或損壞了，從最近一次得到應(yīng)答的數(shù)據(jù)幀 開始的所有幀都進(jìn)行重傳,3. 選擇

7、重傳協(xié)議,接收設(shè)備必須具有排序的邏輯功能. 發(fā)送設(shè)備必須具有查找機(jī)制來幫助它只發(fā)現(xiàn)和選擇需要重傳 的幀. 接收方保存以前接收的各幀. ACK幀的編號(hào)必須指明所接收（或丟失的）那一幀而不是所期待接收的幀號(hào). 窗口大小為 (N+1)/2 (figure 6.17a b ),損壞幀,數(shù)據(jù)幀丟失 (figure 6.18),確認(rèn)幀丟失,特點(diǎn)：只有特定的丟失或損壞幀被重發(fā),6.1.4 HDLC-High-level Data Link Control,配置,正常響應(yīng)方式（NRM） 異步響應(yīng)方式（ARM）,(figure 6.19),是 X.25標(biāo)準(zhǔn)的一部分 面向比特的規(guī)程 支持點(diǎn)到點(diǎn)、多點(diǎn)配置中的半雙工

8、或全雙工模式,1. 幀格式 (figure 6.20),Information frame (I-frame) 信息幀 傳輸用戶數(shù)據(jù)以及與用戶數(shù)據(jù)有關(guān)的控制信息,Supervisory frame (S-frame) 監(jiān)控幀 只用來傳輸控制信息,Unnumbered frame (U-frame) 無序號(hào)幀 保留來為系統(tǒng)管理服務(wù),2. 標(biāo)志域,8比特序列： 01111110 表示幀的開始和結(jié)束(figure 6.21),保持?jǐn)?shù)據(jù)透明性,比特填充 (figure 6.22),3. 地址域,包含了幀的來源或終點(diǎn)的從站點(diǎn)地址(figure 6.23),4. 控制域(figure 6.24) 用來進(jìn)行

9、流量管理,I-Frame,N(S) :描述當(dāng)前發(fā)送幀的序號(hào) N(R) :捎帶的確認(rèn)幀號(hào) P/F：探詢/結(jié)束位,等于1才有效,S-Frame 用來在接收方并沒有數(shù)據(jù)發(fā)送時(shí)返回N（R）值,Code : 流量和控制信息 00 RR(接收方就緒). N(R)字段包含有接收方期待接收的下一幀的序號(hào),01 REJ (拒絕) 用于回退n幀ARQ中,10 RNR (接收方未就緒) 告訴發(fā)送方在收到RR幀之前停止發(fā)送,11 SREJ (選擇拒絕) 選擇重傳ARQ中的否定應(yīng)答幀,U-Frame,Example： （figure 6.25） 00001-SNRM (設(shè)置正常響應(yīng)方式) 11000-SARM (設(shè)置為

10、異步響應(yīng)方式),M4M3M2M1M0 :識(shí)別U-幀的類型及其功能,Example,5. 信息域 I-frame: 用戶數(shù)據(jù) S-frame: 沒有信息域 U-frame: 包含網(wǎng)絡(luò)管理信息,6. 幀校驗(yàn)序列FCS 2-4個(gè)字節(jié)CRC,6.2 Network Layer,將源端的分組經(jīng)各種途徑送到目的端,6.2.1 OSI模型中網(wǎng)絡(luò)層提供的服務(wù),面向連接的服務(wù) 面向無連接的服務(wù),特性: 建立邏輯連接 通信過程包括建立連接，傳輸和釋放連接三個(gè)步驟 分組順序發(fā)送,服務(wù)類型: 可靠的報(bào)文流 可靠的字節(jié)流 不可靠的連接,面向連接的服務(wù),特性: 不需要建立邏輯連接 接收報(bào)文的順序可能和發(fā)送報(bào)文的順序不一樣

11、,服務(wù)類型: 不可靠的數(shù)據(jù)報(bào) (such as E-mail) 需要確認(rèn)的數(shù)據(jù)報(bào) (such as registered mail) 查詢-回答 (如數(shù)據(jù)庫查詢),無連接服務(wù),面向連接的服務(wù): 虛電路子網(wǎng) (figure 6.26),面向無連接的服務(wù): 數(shù)據(jù)報(bào)子網(wǎng) (figure 6.27),數(shù)據(jù)報(bào)子網(wǎng)和虛電路子網(wǎng)的區(qū)別 (figure 6.28),2. 通信子網(wǎng) 子網(wǎng): 硬件 -通信線路和路由器的集合 協(xié)議 - 下3層協(xié)議,6.2.2 IP 協(xié)議,Best-effort (盡力而為): IP 不提供差錯(cuò)檢測或跟蹤功能,IP: 不可靠的面向無連接的數(shù)據(jù)報(bào)協(xié)議，盡可能的傳輸分組。,IP 地址,N

12、ote: 一個(gè)連接了多個(gè)網(wǎng)絡(luò)的設(shè)備具有多個(gè)IP地址 (figure 6.33),IP 數(shù)據(jù)包格式 (figure 6.29),網(wǎng)絡(luò)地址 figure6.32,子網(wǎng),兩層結(jié)構(gòu)（非子網(wǎng)化） (figure 6.34),三層結(jié)構(gòu)（子網(wǎng)化） (figure 6.35),三層 : netid, subnetid（子網(wǎng)識(shí)別號(hào)）, hostid,IP數(shù)據(jù)包找地址的三個(gè)步驟： 找到網(wǎng)絡(luò)地址, 找到子網(wǎng), 找到主機(jī),有子網(wǎng)和無子網(wǎng)的地址區(qū)別 (figure 6.36),掩碼,Example : 分配一個(gè)B類地址，其網(wǎng)絡(luò)地址為128.1.0.0 。我們需要建立254個(gè)子網(wǎng)，每個(gè)子網(wǎng)需要支持254個(gè)主機(jī)。,從IP地

13、址中得到網(wǎng)絡(luò)號(hào)的過程(figure 6.37),Solution ： 128.1.0.0 = 10000000.00000001.00000000.00000000 定義子網(wǎng): 128.1.0.0 = 10000000.00000001.00000000.00000000,定義子網(wǎng)掩碼: 網(wǎng)絡(luò)號(hào)： 10000000.00000001.00000000.00000000 = 128.1.0.0 子網(wǎng)掩碼: 11111111.11111111.11111111.00000000 = 255.255.255.0,254個(gè)子網(wǎng)地址： subnet 1： 10000000.00000001.000000

14、01.00000000 = 128.1.1.0 subnet 2： 10000000.00000001.00000010.00000000 = 128.1.2.0 subnet 3： 10000000.00000001.00000011.00000000 = 128.1.3.0 subnet 254： 10000000.00000001.11111110.00000000 = 128.1.254.0,子網(wǎng) 1的地址范圍:,Subnet 1： 10000000.00000001.00000001.00000000 = 128.1.1.0 Low address： 10000000.0000000

15、1.00000001.00000001 = 128.1.1.1 High address： 10000000.00000001.00000001.11111110 = 128.1.1.254,Notice: IP地址的主機(jī)部分不能全為 “1” 或 “0”,無類尋址,地址是連續(xù)的. 地址的數(shù)量是2的冪 (1,2,4,8). 首地址要能整除地址數(shù)量 . (figure 6.38 ),無類地址的掩碼,對(duì)于IPv4, 一塊地址定義成x.y.z.t /n， 其中 x.y.z.t為地址塊中的一個(gè)地址， /n定義了掩碼。,起始地址可設(shè)置最右邊的32-n位都為0求得，或讓給定地址與 掩碼進(jìn)行與運(yùn)算求得。 最后

16、地址可設(shè)置最右邊的32-n位都為1求得，或讓給定地址與 掩碼的反碼進(jìn)行或運(yùn)算求得。 地址個(gè)數(shù)用232n求得，或通過掩碼的反碼加1求得。,Example : 我們知道一個(gè)給定地址 205.16.37.39/28. 問這塊地址的起始地址、最后地址和地址個(gè)數(shù)。,Solution a. 起始地址為給定地址與掩碼的與運(yùn)算,b. 最后地址為給定地址與掩碼的反碼的或運(yùn)算。,205.16.37.32,205.16.37.47,c. 地址個(gè)數(shù)為掩碼的反碼加.,無類地址的子網(wǎng)劃分,Example: An ISP is granted a block of addresses starting with 190.1

17、00.0.0/16 (65,536 addresses). The ISP needs to distribute these addresses to three groups of customers as follows: a. The first group has 64 customers; each needs 256 addresses. b. The second group has 128 customers; each needs 128 addresses. c. The third group has 128 customers; each needs 64 addre

18、sses. Design the subblocks and find out how many addresses are still available after these allocations.,Solution,Group 1 For this group, each customer needs 256 addresses. This means that 8 (log2 256) bits are needed to define each host. The prefix length is then 32 8 = 24. The addresses are,Group 2

19、 For this group, each customer needs 128 addresses. This means that 7 (log2 128) bits are needed to define each host. The prefix length is then 32 7 = 25. The addresses are,Group 3 For this group, each customer needs 64 addresses. This means that 6 (log264) bits are needed to each host. The prefix l

20、ength is then 32 6 = 26. The addresses are,Number of granted addresses to the ISP: 65,536 Number of allocated addresses by the ISP: 40,960 Number of available addresses: 24,576 (figure 6.39 ),6.2.3 網(wǎng)絡(luò)層控制協(xié)議,網(wǎng)絡(luò)層包括 4個(gè)支持協(xié)議： ARP (Address Resolution Protocol ，地址解析協(xié)議) RARP (Reverse Address Resolution Proto

21、col， 逆地址解析協(xié)議) ICMP (Internet Control Message Protocol， 因特網(wǎng)控制消息協(xié)議) IGMP (Internet Group Multicast Protocol， 因特網(wǎng)組播協(xié)議),ARP (Address Resolution Protocol ),RARP (Reverse Address Resolution Protocol ),RARP 是在已知物理地址的情況下找到IP地址 。,ARP Example: (figure 6.40) information format,Question: 怎樣在已知IP地址的的情況找到硬件地址（物理地

22、址） ?,ICMP (Internet Control Message Protocol ),Each ICMP message is encapsulated（封裝） in an IP packet figure 6.41,ICMP ： 通知發(fā)送方數(shù)據(jù)包是否可以傳遞,6.2.4 路由選擇算法,網(wǎng)絡(luò)層的主要功能： 將分組從源端機(jī)器經(jīng)選定的路由送到目的端機(jī)器,路由選擇算法 ： 負(fù)責(zé)確定所收到的分組應(yīng)傳送的外出路線,特征： 正確性 穩(wěn)定性 簡單性 公平性 健壯性 最優(yōu)性,兩種主要的算法,靜態(tài)路由算法 : 不根據(jù)實(shí)測或估計(jì)的網(wǎng)絡(luò)的當(dāng)前通信量和拓?fù)浣Y(jié)構(gòu)來做路由 選擇，又稱非自適應(yīng)路由算法,動(dòng)態(tài)路由算法

23、 : 根據(jù)拓?fù)浣Y(jié)構(gòu)、通常還有通信量的變化來改變其路由選擇，又稱自適應(yīng)路由算法,距離矢量路由選擇 鏈路狀態(tài)路由選擇,最短路由選擇 基于流量的路由選擇 洪泛法 （擴(kuò)散法）,1. 最優(yōu)化原則,如果路由器J在從路由器I到K的最佳路由上，那么從J到K的最佳線路就會(huì)在同一路由之中,匯集樹： 從所有源端到目的端的最佳路由集合，形成了以目的地為根的樹,目的：為所有路由器找出并使用匯集樹,匯集樹并不唯一(figure 6.43), 靜態(tài)路由算法,2. Shortest Path Routing,主要思想:建立一個(gè)子網(wǎng)圖，圖中的每個(gè)節(jié)點(diǎn)代表一個(gè)路由器，每條弧線表示一條通信線路。為選擇一對(duì)路由器間的路由，算法要在其

24、間找出最短的路徑,測量路徑長度的幾種方法 : 站點(diǎn)數(shù) 地理距離 平均傳輸時(shí)延,Dijkstra 算法: 每個(gè)節(jié)點(diǎn)用從源節(jié)點(diǎn)沿已知最佳路徑到本節(jié)點(diǎn)的距離來標(biāo)注 (figure 6.44 Appendix D),3. 擴(kuò)散法,將收到的每一個(gè)分組，從除了分組到來的線路外的所有輸出 線路上發(fā)出。 會(huì)產(chǎn)生重復(fù)分組,抑制重復(fù)分組的方法： 分組頭中包含站點(diǎn)計(jì)數(shù)器，值為0時(shí)則丟棄 記錄分組擴(kuò)散的路徑，防止第二次擴(kuò)散到已擴(kuò)散的路徑中 選擇性擴(kuò)散 僅發(fā)送到與正確方向接近的那些線路上,1.距離矢量路由選擇,每個(gè)路由器維護(hù)一張表（即向量），表中給出了到每個(gè)目的地的最佳距離和路線。通過與相鄰路由器交換信息來更新表的信息

25、,距離的衡量標(biāo)準(zhǔn): 站點(diǎn)數(shù) 估計(jì)的時(shí)間延遲 路由排隊(duì)的分組估計(jì)總數(shù)載荷 其他, 動(dòng)態(tài)路由算法,Bellman-Ford Algorithm:,設(shè)Cost（A，Z）是從A 到Z的最省路線的費(fèi)用。假設(shè)A與節(jié)點(diǎn)B， C，D有直接相連,則,Example: 假定用延遲來作為度量標(biāo)準(zhǔn)，且路由器知道其相鄰路由器的延遲 (figure 6.45),Typical Protocol: RIP (routing information protocol, 路由信息協(xié)議),兩個(gè)主要問題導(dǎo)致了距離矢量路由選擇算法的消亡,在選擇路由時(shí)，并沒有將線路的帶寬考慮進(jìn)去,算法耗去過多的時(shí)間用于記錄信息(the count-t

26、o-infinity problem).,2.鏈路狀態(tài)路由選擇,發(fā)現(xiàn)它的鄰居節(jié)點(diǎn)，并知道其網(wǎng)絡(luò)地址 -向每條線路發(fā)送特殊的HELLO分組,測量到各鄰居節(jié)點(diǎn)的延遲或開銷 -發(fā)送一個(gè)要求對(duì)方立即響應(yīng)的特殊的ECHO分組以測量來回時(shí)間，可重復(fù)多次取平均值,對(duì)流量（載荷）因素是否考慮(figure 6.46),創(chuàng)建鏈路狀態(tài)分組 (figure 6.47),創(chuàng)建分組的時(shí)間 定期創(chuàng)建 有重大事件時(shí)創(chuàng)建,發(fā)布鏈路狀態(tài)分組(figure 6.48) 基本思想 : 利用擴(kuò)散法發(fā)布分組 分組序號(hào): 控制重復(fù)擴(kuò)散 年齡值:避免分組丟失和無限長地存活下去 傳送標(biāo)志位 :表示分組必須發(fā)向所指示的線路 應(yīng)答標(biāo)志位 :表示

27、應(yīng)向哪兒發(fā)送應(yīng)答,計(jì)算到每個(gè)其他路由器的最短路徑 (Dijkstra algorithm),Typical Protocol: OSPF (open , 開放最短路徑優(yōu)先協(xié)議),6.3 網(wǎng)絡(luò)互聯(lián)設(shè)備,網(wǎng)絡(luò)設(shè)備: 中繼器,網(wǎng)際設(shè)備: 網(wǎng)橋， 路由器，網(wǎng)關(guān),1.中繼器,中繼器是一個(gè)電子設(shè)備，僅僅作用在OSI模型的物理層上,中繼器不以任何方式改變網(wǎng)絡(luò)的功能，僅僅延長網(wǎng)絡(luò)的實(shí)際 距離 (figure 6.49),是一個(gè)再生器而不僅僅是一個(gè)增幅器。 (figure 6.50),在鏈路上的位置影響了差錯(cuò)率,2. 網(wǎng)橋,網(wǎng)橋主要作用在數(shù)據(jù)鏈路層 使不同的網(wǎng)段之間的通信量相互隔離 (figure 6.52)

28、能夠連接不同協(xié)議的LAN (figure 6.51),需要考慮的問題: 幀格式 有效載荷大小 數(shù)據(jù)速率 地址位序 其他問題,3. 路由器,主要工作在OSI模型的網(wǎng)絡(luò)層,Function : 在多個(gè)互聯(lián)網(wǎng)絡(luò)之間中繼包 (figure 6.53 ) 路由選擇 是用于大型網(wǎng)絡(luò)和復(fù)雜的互聯(lián)網(wǎng)絡(luò),4. 網(wǎng)關(guān) 網(wǎng)關(guān)是一個(gè)協(xié)議轉(zhuǎn)換器，用于不同體系結(jié)構(gòu)網(wǎng)絡(luò)的互連 (figure 6.54),6.4 Virtual-Circuit Networks,Faults of X.25: lower data rate and expensive network cost (figure 6.55) providing

29、 a fixed-rate data； large overhead （figure 6.56）,6.4.1 Frame relay,Advantages of frame relay high data rate at lower cost variable length fames with maximum length 9000 bytes operating in physical and data link layers so as to provide services to protocols that already have a network layer (figure 6

30、.57) less overhead for improved transmission media (figure 6.58) allowing bursty data (figure 6.59),Disadvantages: Variable length frames create varying delays Not suitable for sending delay sensitive data such as a real time voice or video,Frame format：（figure 6.60）,Address field：24字節(jié) DLCI： 10 bits

31、 (0-1023), local C/R：1 bit. It is not used by frame relay protocols. EA ：1 bit. (figure 6.61) FECN ：1 bit BECN：1 bit DE：1 bit,Frame relay operation： PVC (permanent virtual circuit)（figure 6.62） SVC ( switched virtual circuit )（figure 6.63）,Congestion control： congestion avoidance using FECN and BECN

32、 to explicitly warn the source and the destination of the presence of congestion (figure 6.64) discarding discarding the frame with DE=1 for serious congestion,6.4.2 Asynchronous Transfer Mode,Optical fiber will allow high-speed interconnection.,The differential size frames in frame relay protocol c

33、reate the varying delay times. (figure 6.65),ATM uses the cell as the basic data unit in combination with switching and multiplexing. A cell is defined as a small fixed-size block of information. ATM uses asynchronous time-division multiplexing. (figure 6.66),ATM architecture（figure 6.67）,A user-to-

34、network interface (UNI) is the interface between a user and an ATM switch.,a network-to-network interface (NNI) is the interface between two ATM switches.,Connection between two end point is accomplished through transmission paths (TPs), virtual paths (VPs), and virtual circuits (VCs). (figure 6.68)

35、,A virtual connection is defined by a pair of numbers: the VPI and the VCI .,ATM layers（figure 6.69） ATM data packet is a cell composed of 53 bytes , including 5 bytes of header and 48 bytes of payload,Physical layer,ATM layer: （figure 6.70） providing routing, traffic management, switching, and mult

36、iplexing service,Header format: 5 bytes (figure 6.71) GFC (generic flow control): 4 bits VPI : 8 bits for UNI cell; 12 bits for NNI cell VCI: 16 bits PT (payload type): 3 bits CLP (cell loss priority): 1 bit HEC (header error correction):16 bit,AAL layer (application adaptation layer) Convergence su

37、blayer (CS) Segmentation and reassembly (SAR),Four different AALs, each for a specific data type AAL1-constant bit-rate stream (figure 6.72) AAL2-variable bit-rate stream (figure 6.73) AAL3/4-conventional packet switching (figure 6.74) AAL5-packets requiring no information from the SAR layer (figure

38、 6.75),6.5 Summary,The task of data link layer is to convert the raw bit stream offered by physical layer into a stream of frames or contrary,Four framing methods are used and the checksum for each frame is computed.,HDLC is a part of X.25 network interface standard and bit- oriented protocol.,Slidi

39、ng window mechanism integrates error control and flow control.,Network layer provides services to transport layer. It can be based on either virtual circuits or datagrams.,Routing algorithms include shortest path routing, flow-based routing , flooding and distance vector routing , link state routing

40、 , etc.,Main function of network layer is Routing and IP routing,IP is defined at network layer. IP is unreliable and connectionless.,IP address uniquely defines connection of a host to its network.,Subnetting allows an additional level of hierarchy in IP address.,ARP(address resolution protocol) fi

41、nds physical address of a device if its IP address is known.,RARP (reverse address resolution protocol) will find a hosts IP address from its physical address.,Network devices is Repeater and Internetwork devices include Bridges, Routers, and Gateways. They operate in the different layer of the OSI

42、model.,Figure 6.1,Principal service provided by data link layer (Relationship between Packets and Frames),Placement of the Data Link Protocol,Figure 6.2,A Character Stream,Figure 6.3,DLE-Data Link Escape（轉(zhuǎn)義字符） STX-Start of Text ETX-End of Text,(a)網(wǎng)絡(luò)層發(fā)出的數(shù)據(jù),（b）經(jīng)數(shù)據(jù)鏈路層填充后的數(shù)據(jù),（c）數(shù)據(jù)傳送給接收方的網(wǎng)絡(luò)層,Figure 6.4,F

43、igure 6.5,Bit Stuffing,Figure 6.6,A Stop-and-Wait Protocol,Stop-and-Wait ARQ, damaged frame,Figure 6.7,Stop-and-Wait ARQ, lost data frame,Figure 6.8,Stop-and-Wait ARQ, lost ACK frame,Figure 6.9,Figure 6.10,Sender Sliding Window for Go-Back-N ARQ,Window size for Go-Back-N ARQ,Figure 6.11,Figure 6.12,

44、Receiver Sliding Window for Go-Back-N ARQ,Figure 6.12,Example with a window of 3 frames,Figure 6.14,Go-back-n, damaged data frame,Figure 6.15,Go-back-n,lost data frame,Figure 6.16,Go-back-n,lost ACK frame,Figure 6.17a,Sliding Window for selective repeat ARQ,Selective Repeat ARQ, window size,Figure 6

45、.17b,Figure 6.18,Selective-reject, lost data frame,Figure 6.19,NRM,ABM,Figure 6.20,HDLC frame types,Figure 6.21,HDLC Flag Field,Figure 6.22,Bit Stuffing and Removal,Figure 6.23,HDLC Address Field,Figure 6.24,HDLC control fields in extended mode,I-Frame,S-Frame,U-Frame,U-frame control command and res

46、ponse,Figure 6.25,Example,Figure 6.26,SVC,Source-to-destination data transfer in a virtual-circuit subnet,Figure 6.27,Datagram,Routing table in a datagram subnet,Figure 6.28,Term,IP Datagram,Figure 6.29,IP address formats,Figure 6.30,Number of blocks and block size in classful IPv4 addressing,In cla

47、ssful addressing, a large part of the available addresses were wasted.,Note,Finding the address class,Figure 6.31,Figure 6.32,Network and Host Addresses,Figure 6.33,A Network with Two Levels of Hierarchy,Figure 6.34,A Network with Three Levels of Hierarchy,Figure 6.35,Addresses in network with or wi

48、thout Subnet,Figure 6.36,Masking,Figure 6.37,Default masks for classful addressing,Example for classless addressing,Figure 6.38,Figure 6.39,ARP,Figure 6.40,Figure 6.41,ICMP encapsulation,The principal ICMP message types,Figure 6.42,(a) A subnet (b) A sink tree for router B,Figure 6.43,hunt shortest

49、path from A to D,Figure 6.44,Figure 6.45,(a) A subnet. (b) Input from A, I, H, K, and the new routing table for J,count-to-infinity problem,Figure 6.46,A subnet in which the East and West parts are connected by two lines,Figure 6.47,(a) A subnet. (b) The link state packets for this subnet.,Figure 6.

50、48,The packet buffer for router B in figure 6.29,Figure 6.49,A Repeater,Figure 6.50,Function of Repeater,Figure 6.51,Function of Bridge,Figure 6.52,Operation of a LAN bridge from 802.11 to 802.3,Figure 6.53,Routers in an Internet,Figure 6.54,A Gateway,Frame Relay versus Pure Mesh T-Line Network,Figu

51、re 6.55,X.25 Traffic,Figure 6.56,Figure 6.57,Comparing Layers in Frame Relay and X.25,Frame Relay Traffic,Figure 6.58,Fixed-Rate versus Bursty Data,Figure 6.59,Frame Relay Frame,Figure 6.60,Three address formats,Figure 6.61,Figure 6.62,PVC,SVC Setup and Release,Figure 6.63,Congestion avoidance,Figur

52、e 6.64,Figure 6.65,Multiplexing using different frame sizes,ATM Multiplexing using cells,Figure 6.66,Figure 6.67,Architecture of an ATM network,Connection identifiers,Figure 6.68,ATM layers,Figure 6.69,Figure 6.70,ATM Layer,Figure 6.71,ATM Header,Figure 6.72,AAL1,Figure 6.73,AAL2,Figure 6.74,AAL3/4,Figure 6.75,AAL5,/* Protocol (utopia烏托邦) */ typedef enum frame_arrival event_type; #include protocol.h void sender1(void) frame s; /* buffer for an outbound