Network Layer Overview: The Data Link Layer � network layer services � virtual circuit and datagram Goals: networks Today Our goals: � what’s inside a router? � understand principles � link layer services � understand principles behind network layer � IP: Internet Protocol � error detection, correction behind data link layer services: � IPv4 datagram format services: � multiple access protocols and � forwarding � IPv4 addressing LANs � routing (path selection) � error detection, � ICMP � dealing with scale correction � link layer addressing, ARP, � IPv6 DHCP � how a router works � routing algorithms � sharing a broadcast � advanced topics: IPv6, � Link state channel: multiple access Next time multicast � Distance Vector � link layer addressing � Ethernet � instantiation and � Hierarchical routing implementation in the � reliable data transfer, � hubs and switches � routing in the Internet Internet flow control: done! � PPP � RIP � instantiation and � OSPF implementation of various � BGP link layer technologies � broadcast and multicast routing 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Link Layer: setting the context Link Layer: setting the context � two physically connected devices: � host-router, router-router, host-host � unit of data: frame M application transport H t M network network M H n H t data link protocol link link H l H n H t M H l H n H t M physical physical frame phys. link adapter card 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Link Layer Services Link Layer Services (more) � Flow Control: � Framing, link access: � pacing between adjacent sending and receiving nodes � encapsulate datagram into frame, adding header, trailer � channel access if shared medium, � Error Detection : � ‘physical addresses’ used in frame headers to identify � errors caused by signal attenuation, noise. source, dest � receiver detects presence of errors: • different from IP address! � Reliable delivery between two physically connected • signals sender for retransmission or drops frame devices: � Error Correction: � we learned how to do this already (chapter 3)! � receiver identifies and corrects bit error(s) without � seldom used on low bit error link (fiber, some twisted resorting to retransmission pair) � Half-duplex and full-duplex � wireless links: high error rates � with half duplex, nodes at both ends of link can transmit, • Q: why both link-level and end-end reliability? but not at same time 26/4-07 26/4-07 Datakommunikation - Jonny Pettersson, UmU Datakommunikation - Jonny Pettersson, UmU 1
Adaptors Communicating Error Detection datagram EDC= Error Detection and Correction bits (redundancy) rcving link layer protocol D = Data protected by error checking, may include header fields node sending node frame frame • Error detection not 100% reliable! • protocol may miss some errors, but rarely adapter adapter • larger EDC field yields better detection and correction � link layer implemented in � sending side: “adaptor” (aka NIC) � encapsulates datagram in a frame � Ethernet card, PCMCIA card, 802.11 card � adds error checking bits, rdt, flow control, etc. � typically includes: RAM, DSP (Digital Signal Processing) � receiving side chips, host bus interface, � looks for errors, rdt, flow and link interface control, etc � adapter is semi-autonomous � extracts datagram, passes to � link & physical layers rcving node 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Parity Checking Checksumming: Cyclic Redundancy Check � view data bits, D, as a binary number Two Dimensional Bit Parity : Single Bit Parity: � choose r+1 bit pattern (generator), G Detect and correct single bit errors Detect single bit errors � goal: choose r CRC bits, R, such that <D,R> exactly divisible by G (modulo 2) � � receiver knows G, divides <D,R> by G. If non-zero remainder: error detected! � can detect all burst errors less than r+1 bits and any odd number of bit errors � a burst of length greater than r+1 bits is detected with probability 1-0.5 r � widely used in practice (ATM, HDCL) 0 0 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Multiple Access protocols Ideal Mulitple Access Protocol � single shared communication channel Broadcast channel of rate R bps � two or more simultaneous transmissions by nodes: 1. When one node wants to transmit, it can send at interference rate R � only one node can send successfully at a time � multiple access protocol: 2. When M nodes want to transmit, each can send at � distributed algorithm that determines how nodes share average rate R/M channel, i.e., determine when node can transmit 3. Fully decentralized: � communication about channel sharing must use channel itself! � no special node to coordinate transmissions � what to look for in multiple access protocols: � no synchronization of clocks, slots • synchronous or asynchronous 4. Simple • information needed about other nodes • robustness (e.g., to channel errors) • performance 26/4-07 26/4-07 Datakommunikation - Jonny Pettersson, UmU Datakommunikation - Jonny Pettersson, UmU 2
MAC Protocols: a taxonomy Channel Partitioning MAC protocols: TDMA (MAC – Media Access Control) TDMA: time division multiple access Three broad classes: � access to channel in "rounds" � Channel Partitioning � each station gets fixed length slot (length = pkt � divide channel into smaller “pieces” (time slots, trans time) in each round frequency, code) � allocate piece to node for exclusive use � unused slots go idle � Random Access � example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6 � channel not divided, allow collisions idle � “recover” from collisions � “Taking turns” � tightly coordinate shared access to avoid collisions Goal: efficient, fair, simple, decentralized 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Channel Partitioning (CDMA) Channel Partitioning MAC protocols: FDMA FDMA: frequency division multiple access CDMA (Code Division Multiple Access) � channel spectrum divided into frequency bands � unique “code” assigned to each user; ie, code set � each station assigned fixed frequency band partitioning � unused transmission time in frequency bands go idle � used mostly in wireless broadcast channels � example: 6-station LAN, 1,3,4 have pkt, frequency (cellular, satellite,etc) bands 2,5,6 idle � all users share same frequency, but each user has t i m e own “chipping” sequence (ie, code) to encode data frequency bands � allows multiple users to “coexist” and transmit simultaneously with minimal interference (if codes are “orthogonal”) � more later… 26/4-07 Datakommunikation - Jonny Pettersson, UmU 26/4-07 Datakommunikation - Jonny Pettersson, UmU Random Access protocols Slotted ALOHA � When node has packet to send Assumptions Operation � transmit at full channel data rate R � all frames same size � when node obtains fresh � no a priori coordination among nodes frame, it transmits in next � time is divided into � two or more transmitting nodes -> “collision”, slot equal size slots = time � random access MAC protocol specifies: to transmit 1 frame � no collision, node can send � how to detect collisions new frame in next slot � nodes start to transmit � how to recover from collisions (e.g., via delayed frames only at � if collision, node retransmissions) beginning of slots retransmits frame in each � Examples of random access MAC protocols: subsequent slot with prob. � nodes are synchronized � slotted ALOHA p until success � if 2 or more nodes � ALOHA transmit in slot, all � CSMA and CSMA/CD nodes detect collision 26/4-07 26/4-07 Datakommunikation - Jonny Pettersson, UmU Datakommunikation - Jonny Pettersson, UmU 3
Recommend
More recommend