Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Chapter 3 Transport Layer Modified by John Copeland Georgia Tech for use in ECE3600 A note on the use of these ppt slides: We’re making these slides freely available to all (faculty, students, readers). They’re in PowerPoint form so you can add, modify, and delete slides (including this one) and slide content to suit your needs. They obviously represent a lot of work on our part. In return for use, we only ask the following: If you use these slides (e.g., in a class) in substantially unaltered form, that you mention their source (after all, we’d like people to use our book!) If you post any slides in substantially unaltered form on a www site, that you note that they are adapted from (or perhaps identical to) our slides, and note our copyright of this material. Computer Networking: A Top Down Approach Featuring the Internet, 5th edition. Jim Kurose, Keith Ross Addison-Wesley, July 2009. Thanks and enjoy! JFK/KWR All material copyright 1996-2009 J.F Kurose and K.W. Ross, All Rights Reserved Transport Layer 10-4-2014 3-1 Chapter 3: Transport Layer Our goals: understand principles behind transport layer services: multiplexing/demultiplexing reliable data transfer flow control congestion control learn about transport layer protocols in the Internet: UDP: connectionless transport TCP: connection-oriented transport TCP congestion control Transport Layer 3-2 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-3 Transport services and protocols provide logical communication between app processes running on different hosts transport protocols run in end systems send side: breaks app messages into segments, passes to network layer receive side: reassembles segments into messages, passes to app layer more than one transport protocol available to apps Internet: TCP and UDP application transport network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical application transport network data link physical Transport Layer 3-4 Transport vs. network layer network layer: logical communication between hosts transport layer: logical communication between processes* relies on, enhances, network layer services * running on different hosts Household analogy: 12 kids sending letters to 12 kids processes = kids app messages = letters in envelopes hosts = houses transport protocol = Ann and Bill network-layer protocol = postal service Transport Layer 3-5 Internet transport-layer protocols reliable, in-order delivery (TCP) congestion control flow control connection setup unreliable, unordered delivery: UDP no-frills extension of “best-effort” IP services not available: delay guarantees bandwidth guarantees application transport network data link physical network data link physical network data link physical network data link physical network data link physical network data link physical application transport network data link physical Transport Layer 3-6 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-7 Multiplexing/demultiplexing Multiplexing at send host: gathering data from multiple sockets, enveloping data with header (later used for demultiplexing) Demultiplexing at rcv host: delivering received segments to correct socket = socket application transport network link = process P3 P1 P1 application transport network P2 P4 application transport network link link physical host 1 physical host 2 physical host 3 Transport Layer 3-8 How demultiplexing works host receives IP datagrams each datagram has source IP address, destination IP address each datagram carries 1 transport-layer segment each segment has source, destination port number host uses IP addresses & port numbers to direct segment to appropriate socket 32 bits source port # dest port # other header fields application data (message) TCP and UDP segment format Transport Layer 3-9 Connectionless demultiplexing Create sockets with port numbers: DatagramSocket mySocket1 = new DatagramSocket(12534); DatagramSocket mySocket2 = new DatagramSocket(12535); UDP socket identified by two-tuple (plus "UDP", it is a 3-tuple): (For incoming segment: destination IP address, destination port number) When host receives UDP segment: checks destination port number in segment directs UDP segment to socket with that port number IP datagrams with different source IP addresses and/or source port numbers directed to same UDP socket This means a single process thread can handle all UDP packets coming to a host that have the same local (destination) UDP port number. This is not the case for TCP (an addition thread is needed for each remote IP address and TCP port pair). Transport Layer 3-10 Connectionless (UDP) demux (cont) DatagramSocket serverSocket = new DatagramSocket(6428); P2 SP: 6428 SP: 6428 DP: 9157 DP: 5775 SP: 9157 client IP: A P1 P1 P3 DP: 6428 SP provides “return address” SP: 5775 server IP: C DP: 6428 Client IP:B Same incoming UDP destination port, delivered to same process, the one that opened that socket. Transport Layer 3-11 Connection-oriented demux TCP socket identified by a 4-tuple (+ "TCP" -> a 5-tuple): source IP address source port number dest IP address dest port number recv host TCP layer uses all four values to direct segment to appropriate socket Server host may support many simultaneous TCP sockets: each TCP socket identified by its own 4-tuple (5-tuple) Web servers have different sockets for each connecting client non-persistent HTTP will have different socket (different client port number) for each request Transport Layer 3-12 Connection-oriented (TCP) demux (cont) Different TCP source port from P5 and P6 on client C -> different process (thread) (P2, P3 on server B) P1 P4 P5 P2 P6 P1P3 SP: 5775 DP: 80 S-IP: C D-IP:B server IP: A SP:5776 DP: 80 S-IP: C D-IP:A SP: 5777 client IP: C DP: 80 S-IP: C D-IP:B server IP:B Transport Layer 3-13 Connection-oriented (TCP) demux: Threaded Web Server P1 Listening Socket SP: * DP: 80 Different TCP source port -> different sub-process (thread) * DP: 80 D-IP:C S-IP: B D-IP:C * = "any" SP: 9157 DP: 80 S-IP: A D-IP:C P1P3 SP: 5775 S-IP: * client IP: A P2 P4 SP: 9157 server IP: C DP: 80 S-IP: B D-IP:C Client IP:B thread, spawned or "forked" from main process, manages each connected socket. Transport Layer 3-14 Active Socket states: ESTABLISHED or LISTEN (netstat) mc:/Users/copeland root# netstat -f inet -al Active Internet connections Proto Recv-Q Send-Q Local Address (IP.port) Foreign Address tcp4 0 0 mc.55185 www.suntrust.com.https tcp4 0 48 mc.ssh home.49901 tcp4 0 0 *.9503 *.* tcp4 0 0 localhost.ipp *.* tcp4 0 0 *.afpovertcp *.* tcp4 0 0 *.timbuktu *.* tcp4 0 0 *.ssh *.* udp4 0 0 *.mdns *.* udp4 0 0 *.52521 *.* udp4 0 0 *.9912 *.* udp4 0 0 *.ipp *.* udp4 0 0 localhost.49156 localhost.1022 udp4 0 0 localhost.49155 localhost.1022 udp4 0 0 localhost.1022 *.* udp4 0 0 localhost.1023 *.* udp4 0 0 *.timbuktu-srv3 *.* udp4 0 0 *.timbuktu *.* udp4 0 0 *.* *.* udp4 0 0 mc.ntp *.* udp4 0 0 localhost.ntp *.* udp4 0 0 *.ntp *.* udp4 0 0 *.tftp *.* (state) ESTABLISHED ESTABLISHED LISTEN LISTEN LISTEN LISTEN LISTEN Well Known Ports (from /etc/services): ssh = 22, mdns = 5353, timbuktu = 407, timbuktu-srv3 = 1419, ntp = 123, tftp = 69, ipp = 631 Host names (from /etc/hosts): mc, home (from DNS) www.suntrust.com Transport Layer 3-15 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-16 UDP: User Datagram Protocol [RFC 768] “no frills,” “bare bones” Internet transport protocol “best effort” service, UDP segments may be: lost delivered out of order to app connectionless: no handshaking between UDP sender, receiver each UDP segment handled independently of others Why is there a UDP? no connection establishment (which can add delay) simple: no connection state at sender, receiver small segment header no congestion control: UDP can blast away as fast as desired Transport Layer 3-17 UDP: more often used for streaming multimedia apps loss tolerant rate sensitive Length, in bytes of UDP segment, including other UDP uses header DNS (name lookup) SNMP (network management) NTP (network time protocol) to get reliable transfer over UDP: add reliability at application layer application-specific error recovery! 32 bits source port # dest port # length checksum Application data (message) UDP segment format Transport Layer 3-18 UDP checksum Goal: detect “errors” (e.g., flipped bits) in transmitted segment Sender: Receiver: treat segment contents compute checksum of as sequence of 16-bit integers checksum: addition (1’s complement sum) of segment contents sender puts checksum value into UDP checksum field received segment check if computed checksum equals checksum field value: NO - error detected YES - no error detected. But maybe errors nonetheless? More later …. Transport Layer 3-19 Internet Checksum Example Note When adding numbers, a carryout from the most significant bit needs to be added to the result Example: add two 16-bit integers (1's compliment) 1 1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 wraparound 1 1 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1 sum 1 1 0 1 1 1 0 1 1 1 0 1 1 1 1 0 0 checksum 1 0 1 0 0 0 1 0 0 0 1 0 0 0 0 1 1 Transport Layer 3-20 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-21 Principles of Reliable data transfer important in app., transport, link layers top-10 list of important networking topics! characteristics of unreliable channel will determine complexity of reliable data transfer protocol (rdt) Transport Layer 3-22 Principles of Reliable data transfer important in app., transport, link layers top-10 list of important networking topics! characteristics of unreliable channel will determine complexity of reliable data transfer protocol (rdt) Transport Layer 3-23 Principles of Reliable data transfer important in app., transport, link layers top-10 list of important networking topics! characteristics of unreliable channel will determine complexity of reliable data transfer protocol (rdt) Transport Layer 3-24 Reliable data transfer: getting started rdt_send(): called from above, (e.g., by app.). Passed data to deliver to receiver upper layer send side udt_send(): called by rdt, to transfer packet over unreliable channel to receiver deliver_data(): called by rdt to deliver data to upper receive side rdt_rcv(): called when packet arrives on rcv-side of channel Transport Layer 3-25 Reliable data transfer: getting started We’ll: incrementally develop sender, receiver sides of reliable data transfer protocol (rdt) consider only unidirectional data transfer but control info will flow on both directions! use finite state machines (FSM) to specify sender, receiver state: when in this “state” next state uniquely determined by next event state 1 event causing state transition actions taken on state transition event actions state 2 Transport Layer 3-26 Rdt1.0: reliable transfer over a reliable channel underlying channel perfectly reliable no bit errors Event causes change of state Action that occurs no loss of packets separate FSMs for sender, receiver: sender sends data into underlying channel receiver read data from underlying channel Wait for call from above rdt_send(data) packet = make_pkt(data) udt_send(packet) sender Wait for call from below rdt_rcv(packet) extract (packet,data) deliver_data(data) receiver Transport Layer 3-27 Rdt2.0: channel with bit errors underlying channel may flip bits in packet checksum to detect bit errors the question: how to recover from errors: acknowledgements (ACKs): receiver explicitly tells sender that pkt received OK negative acknowledgements (NAKs): receiver explicitly tells sender that pkt had errors sender retransmits pkt on receipt of NAK new mechanisms in rdt2.0 (beyond rdt1.0): error detection receiver feedback: control msgs (ACK,NAK) rcvr->sender Transport Layer 3-28 rdt2.0: FSM specification rdt_send(data) snkpkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for Wait for call from ACK or udt_send(sndpkt) above NAK rdt_rcv(rcvpkt) && isACK(rcvpkt) L sender receiver rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) Wait for call from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) Transport Layer 3-29 rdt2.0: operation with no errors rdt_send(data) snkpkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for Wait for call from ACK or udt_send(sndpkt) above NAK rdt_rcv(rcvpkt) && isACK(rcvpkt) L rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) Wait for call from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) Transport Layer 3-30 rdt2.0: error scenario rdt_send(data) snkpkt = make_pkt(data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) Wait for Wait for call from ACK or udt_send(sndpkt) above NAK rdt_rcv(rcvpkt) && isACK(rcvpkt) L rdt_rcv(rcvpkt) && corrupt(rcvpkt) udt_send(NAK) Wait for call from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) Transport Layer 3-31 rdt2.0 has a fatal flaw! What happens if ACK/NAK corrupted? sender doesn’t know what happened at receiver! can’t just retransmit: possible duplicate Handling duplicates: sender retransmits current pkt if ACK/NAK garbled sender adds sequence number to each pkt receiver discards (doesn’t deliver up) duplicate pkt stop and wait Sender sends one packet, then waits for receiver response Transport Layer 3-32 rdt2.1: sender, handles garbled ACK/NAKs rdt_send(data) sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || Wait for Wait for isNAK(rcvpkt) ) ACK or call 0 from udt_send(sndpkt) NAK 0 above rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt) L rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isNAK(rcvpkt) ) udt_send(sndpkt) L Wait for ACK or NAK 1 Wait for call 1 from above rdt_send(data) sndpkt = make_pkt(1, data, checksum) udt_send(sndpkt) Transport Layer 3-33 rdt2.1: receiver, handles garbled ACK/NAKs rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq0(rcvpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt) sndpkt = make_pkt(NAK, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && not corrupt(rcvpkt) && has_seq1(rcvpkt) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) sndpkt = make_pkt(NAK, chksum) udt_send(sndpkt) Wait for 0 from below Wait for 1 from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq1(rcvpkt) rdt_rcv(rcvpkt) && not corrupt(rcvpkt) && has_seq0(rcvpkt) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) Transport Layer 3-34 rdt2.1: discussion Sender: seq # added to pkt two seq. #’s (0,1) will suffice. Why? must check if received ACK/NAK corrupted twice as many states state must “remember” whether “current” pkt has 0 or 1 seq. # Receiver: must check if received packet is duplicate state indicates whether 0 or 1 is expected pkt seq # note: receiver can not know if its last ACK/NAK received OK at sender Transport Layer 3-35 rdt2.2: a NAK-free protocol same functionality as rdt2.1, using ACKs only instead of NAK, receiver sends ACK for last pkt received OK receiver must explicitly include seq # of pkt being ACKed [say “0”] duplicate ACK at sender [two ACKs for “0”] results in same action as NAK: retransmit current pkt [“1”] Transport Layer 3-36 rdt2.2: sender, receiver fragments rdt_send(data) sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || Wait for Wait for isACK(rcvpkt,1) ) ACK call 0 from 0 udt_send(sndpkt) above sender FSM fragment rdt_rcv(rcvpkt) && (corrupt(rcvpkt) || has_seq1(rcvpkt)) Wait for 0 from below sndpkt = make_pkt(ACK0, chksum) udt_send(sndpkt) “Duplicate ACK” rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,0) receiver FSM fragment L rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq1(rcvpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK1, chksum) udt_send(sndpkt) Transport Layer 3-37 rdt3.0: channels with errors and loss New assumption: underlying channel can also lose packets (data or ACKs) checksum, seq. #, ACKs, retransmissions will be of help, but not enough Approach: sender waits “reasonable” amount of time for ACK retransmits if no ACK received in this time if pkt (or ACK) just delayed (not lost): retransmission will be duplicate, but use of seq. #’s already handles this receiver must specify seq # of pkt being ACKed requires countdown timer Transport Layer 3-38 rdt3.0 sender rdt_send(data) sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) start_timer rdt_rcv(rcvpkt) L Wait for call 0 from above rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,1) timeout udt_send(sndpkt) start_timer rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isACK(rcvpkt,0) ) L Wait for ACK0 timeout udt_send(sndpkt) start_timer rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,0) stop_timer stop_timer L rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isACK(rcvpkt,1) ) Wait for ACK1 Wait for call 1 from above rdt_send(data) rdt_rcv(rcvpkt) L sndpkt = make_pkt(1, data, checksum) udt_send(sndpkt) start_timer Transport Layer 3-39 rdt3.0 in action Transport Layer 3-40 rdt3.0 in action Transport Layer 3-41 Review rdt3.0 ACK Packet Lets Sender know that data has been received. Sequence Number In Header - Lets Receiver know where to place data in buffer (to avoid duplication and gaps) Acknowledgement Number In ACK - Lets Sender know which data has been received. (so it knows what to resend) Retransmit Timer Keeps connection from freezing because of a lost packet. Transport Layer 3-42 Performance of rdt3.0 rdt3.0 works, but performance stinks example: 1 Gbps link, 15 ms e-e prop. delay, 1KB packet: Ttransmit = L (packet length in bits) 8kb/pkt = = 0.008 millisec R (transmission rate, bps) 10**9 b/sec U sender: utilization – fraction of time sender busy sending 1KB pkt every 30 msec -> 33kB/sec thruput over 1 Gbps link network protocol limits use of physical resources! Transport Layer 3-43 rdt3.0: stop-and-wait operation sender receiver first packet bit transmitted, t = 0 last packet bit transmitted, t = L / R RTT first packet bit arrives last packet bit arrives, send ACK ACK arrives, send next packet, t = RTT + L / R U is the “channel utilization factor”, or “efficiency” Transport Layer 3-44 Pipelined protocols Pipelining: sender allows multiple, “in-flight”, yetto-be-acknowledged pkts range of sequence numbers must be increased buffering at sender and/or receiver Two generic forms of pipelined protocols: go-Back-N, selective repeat Transport Layer 3-45 Pipelining: increased utilization first packet bit transmitted, t=0 last bit transmitted, t = L / R sender RTT receiver first packet bit arrives last packet bit arrives, send ACK 1 last bit of 2nd packet arrives, send ACK 2 last bit of 3rd packet arrives, send ACK 3 ACK arrives, send next packet, t = RTT + L / R Increase utilization by a factor of 3 ! If 3L/R < RTT and Window = 3L, then Throughput Rt = R*U = (Window/RTT) Transport Layer 3-46 Go-Back-N Sender: k-bit seq # in pkt header “window” of up to N, consecutive unack’ed pkts allowed ACK(n): ACKs all pkts up to, including seq # n - “cumulative ACK” may receive duplicate ACKs (see receiver) timer for each in-flight pkt timeout(n): retransmit pkt n and all higher seq # pkts in window Transport Layer 3-47 Go Back N: Receiver ACK-only: always send ACK for correctly-received pkt with highest in-order seq # may generate duplicate ACKs need only remember expectedseqnum out-of-order pkt: discard (don’t buffer) -> no receiver buffering! Re-ACK packet with the highest in-order seq # Transport Layer 3-48 GBN in action Transport Layer 3-49 Selective Repeat receiver individually acknowledges all correctly received pkts buffers pkts, as needed, for eventual in-order delivery to upper layer sender only resends pkts for which ACK not received sender timer for each unACKed pkt sender window N consecutive seq #’s again limits seq #s of sent, unACKed pkts Transport Layer 3-50 Selective repeat: sender, receiver windows send_base moves forward to first notack’ed packet rcv_base moves forward to first nottransferred packet (TCP buffers all, but ACK’s first missing byte, = rcv-base) Transport Layer 3-51 Selective repeat sender data from above : receiver pkt n in [rcvbase, rcvbase+N-1] if next available seq # in send ACK(n) timeout(n): in-order: deliver (also window, send pkt resend pkt n, restart timer ACK(n) in [sendbase,sendbase+N]: mark pkt n as received if n smallest unACKed pkt, advance window base to next unACKed seq # out-of-order: buffer deliver buffered, in-order pkts), advance window to next not-yet-received pkt pkt n in [rcvbase-N,rcvbase-1] ACK(n) (already received & delivered) otherwise: Ignore Transport Layer 3-52 Selective repeat in action TCP ack1 ack2 ACK(2nd) ack2 (3rd) ack2 (4th) On 4th ack2 The sender resends only Pkt2, Then continues Sender must keep a list of pkt’s ACKed Transport Layer 3-53 Selective repeat: dilemma Example: seq #’s: 0, 1, 2, 3 window size=3 receiver sees no difference in two scenarios! incorrectly passes duplicate data as new in (a) Q: what relationship between seq # size and window size? Transport Layer 3-54 Reliable Data Transport Problem Solution Packet may arrive with errors. Add checksum, CRC, or hash. Packet may not arrive. Receiver sends “ACK” back. If ACK not received, packet re-sent. Sender may wait forever for ACK. Timeout timer added to sender. ACK may not arrive, dup. sent. Add sequence no.s to detect dups. Packets may arrive out-of-order. Buffer packets to rearrange order. Inefficient to send one pkt per RT Have a “window” to send before ACK (pipelining). Missing packet early in window. “Go-Back-N” to last in-order packet. “Go-Back-N” inefficient. “Selective Repeat” to fill in gaps only. ---- Also in TCP --- ---- Packets may be different sizes. Sequence number for each byte. Slow down when network “Slow-Start”, or "Multiplicative congested (as detected by RTO or Decrease" to reduce transmit window. triple duplicate ACKs. Know when receiver buffer will be full. Receiver includes “space left” in every ACK. Transport Layer 3-55 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-56 TCP: Overview point-to-point: one sender, one receiver reliable, in-order byte steam: no “message boundaries” pipelined: TCP congestion and flow control set window size send & receive buffers socket door application writes data application reads data TCP send buffer TCP receive buffer RFCs: 793, 1122, 1323, 2018, 2581 full duplex data: bi-directional data flow in same connection MSS: maximum segment size connection-oriented: handshaking (exchange of control msgs) init’s sender, receiver state before data exchange flow controlled: sender will not socket door overwhelm receiver segment Transport Layer 3-57 TCP segment structure 32 bits URG: urgent data (generally not used) ACK: ACK # valid PSH: push data now (generally not used) RST, SYN, FIN: connection estab (setup, teardown commands) Internet checksum (as in UDP) source port # dest port # sequence number acknowledgement number head not UA P R S F len used checksum Receive window Urg data pnter Options (variable length) application data (variable length) counting by bytes of data (not segments!) # bytes rcvr willing to accept Seg. Size = IP Size – IP Header Size Data Size = IP Size – IP Header Size - TCP Header Size Transport Layer 3-58 TCP seq. #’s and ACKs Seq. #’s: byte stream “number” of first byte in segment’s data ACKs: seq # of next byte expected from other side cumulative ACK Q: how receiver handles out-of-order segments A: TCP spec doesn’t say, - up to implementor Host A User types ‘C’ Host B host ACKs receipt of ‘C’, echoes back ‘C’ host ACKs receipt of echoed ‘C’ simple telnet scenario time Transport Layer 3-59 TCP Round Trip Time and Timeout Q: how to set TCP timeout value? longer than RTT but RTT varies if too short: premature timeout and unnecessary retransmissions if too long: slow reaction to segment loss Q: how to estimate RTT? SampleRTT: measured time from segment transmission until ACK receipt ignore measurement if packet was retransmitted. SampleRTT will vary, we want the estimated RTT “smoother” average several recent measurements, not just current SampleRTT by using a "running average. Transport Layer 3-60 TCP Round Trip Time and Timeout EstimatedRTT = (1- )*EstimatedRTT + *SampleRTT This is like a computer instruction, rather than an algebra. It may be clearer to write: EstimatedRTT[new] = (1-)* EstimatedRTT[old] + *SampleRTT[new] where "old" and "new" are array indices, and new = old + 1 Exponential weighted moving average influence of past sample decreases exponentially fast typical value: = 0.125 Transport Layer 3-61 Example RTT estimation: Transport Layer 3-62 TCP Round Trip Time and Timeout Setting the retransmission timeout, RTO: use EstimatedRTT plus “safety margin” larger variation in EstimatedRTT -> larger safety margin use a running average, DevRTT, to estimate how much SampleRTT deviates from EstimatedRTT: DevRTT[new] = (1-)* DevRTT[old] + * |SampleRTT[new]-EstimatedRTT[old*]| (typically, = 0.25) Note the absolute bars,|...| Then set a new timeout interval, RTO: Timeout Interval: RTO = EstimatedRTT + 4 * DevRTT *Note - "old" value used, not "new" value. Transport Layer 3-63 A[i] is a exponentially-decaying (1-c) weighed average of x[i], x[i-1], x[i-2], … A[i] = (1-c) * A[ i – 1 ] + c * x[ i ] A[ i - 1] = (1-c) * A[ i – 2 ] + c * x[ i - 1 ] A[ i - 2] = (1-c) * A[ i – 3 ] + c * x[ i - 2 ] ... A[i] = c * { x[ i ] + (1-c) * x[ i – 1 ] + + (1-c)^2 * x[ i – 2 ] + … } Advantages: Only the last value of A[i] needs to be remembered. More recent values of x[i] are more heavily weighted. 0.25 0.20 c=0.25 0.15 c=0.125 0.10 0.05 0.00 1 2 3 4 5 6 7 8 9 10 64 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-65 TCP reliable data transfer TCP creates rdt service on top of IP’s unreliable service Pipelined segments Window is in bytes Cumulative acks* TCP uses single Retransmissions are triggered by: timeout events (GoBack-N) 3 duplicate acks (like Selective Repeat*) retransmission timer* • Differs from Selective Repeat. If a segment is missing, the receiver keeps sending the same ACK number (the first missing byte). • After 3 duplicate ACKs, the sender sends only the missing segment. Transport Layer 3-66 Maximum Segment Size (MSS), in bytes The initial segments (the SYN and SYN-ACK) contain the MSS in an option field. It stays constant after this. This tells the other host the maximum size of a segment that can be handled by their local network (without fragmentation). Examples, one host may say it's MSS value is 1400, the other may say it's MSS value is 1420. Since segments have to transverse both local networks, the smaller MSS value is used for the connection. TCP rules, involving Window sizes, are in number of MSS (bytes), not number of segments. For simplification, examples may say "the host is sending maximum size segments," so that 1 MSS = 1 segment. Sometimes this is implied without being stated in problems. MSS includes the TCP header bytes (20 to 64) and data bytes, but not the IP header bytes (20). Since Ethernet and WiFi limit datagram size to 1500 bytes, MSS is never larger than 1480 bytes when either host is on a LAN. 3-67 TCP ACK generation [RFC 1122, RFC 2581] Event at Receiver TCP Receiver action Arrival of in-order segment with expected seq #. All data up to expected seq # already ACKed Delayed ACK. Wait up to 500ms for next segment. If no next segment, send ACK [don’t ACK every segment] Arrival of in-order segment with expected seq #. One other segment has ACK pending Immediately send single cumulative ACK, ACKing both in-order segments [ACK every other segment] Arrival of out-of-order segment higher-than-expect seq. # . Gap detected Immediately send duplicate ACK, indicating seq. # of next expected byte [1st byte in gap] Arrival of segment that partially or completely fills gap Immediate send ACK for highest continuous byte. If segment does not start at lower end of gap, it will be another duplicate ACK. Transport Layer 3-68 Two windows, separate limits on bytes sent (Flow Control & Congestion Control). RcvrWin in each TCP header received (room in receiver buffer) CongWin calculated by sender (in response to perceived congestion). Sender Buffer ACK 92, Window 40 received 92 100 SendBase is the lowest unacknowledged byte no., and it is also the highest ACK no. received. 120 + CongWin (30) SendBase = 92 + RcvrWin (40) ACK 120, Window 20 received 122 132 + CongWin (30) SendBase+ RcvrWin (20) = 120 140 160 Sender can not send bytes beyond either window. MSS limits bytes in a segment, constant after SYN / SYN-ACK. Transport Layer 3-69 TCP Flow control LastByte InBuffer LastByte ACKed <- byte no. Rcvr advertises spare room by including value of RcvWindow in every segment (TCP header) Sender limits data to RcvWindow + ACK guarantees receive buffer doesn’t overflow higher byte no.s Receiver-Window = spare room in buffer = LastByteInBuffer LastByteACKed flow control sender won’t overflow receiver’s buffer by transmitting too fast Transport Layer 3-70 TCP sender events: RTO (timeout) data rcvd from app: Create segment with seq # seq # is byte-stream number of first data byte in segment start RTO timer if not already running (think of timer as being for oldest un-ACKed segment) expiration interval: calculate new value for TimeOutInterval(RTO) timeout: retransmit segment that caused timeout (Go-Back-N) RTO -> 2*RTO temporarily CongWin = 1 MSS (maximum segment size, e.g. 1420 bytes)* restart RTO timer ACK rcvd: If acknowledges previously unACKed segments * CongWin has become very small. It can double in size every RTT, but it would take 6 RTT to get back to 64 MSS (Slow Start mode). update what is known to be ACKed (move SendBase to ACK#) CongWin =CongWin + MSS RTO -> original value (from 2X) start RTO timer if there are outstanding segments Transport Layer 3-71 TCP: retransmission scenarios Host A X loss SendBase = 100 SendBase = 120 SendBase = 100 time SendBase = 120 lost ACK scenario Host B Seq=92 timeout Host B Seq=92 timeout timeout Host A time premature timeout Transport Layer 3-72 TCP retransmission scenarios (more) Host A Host B timeout Sender Buffer 92 X 120 + CongWin loss SendBase SendBase = 120 100 + RcvrWin + CongWin + RcvrWin’ SendBase time Cumulative ACK scenario Sender can not send bytes beyond either window. Every ACK has updated value of RcvrWin Transport Layer 3-73 Fast Retransmit Time-out period often relatively long: long delay before resending lost packet Detect lost segments via duplicate ACKs. Sender often sends many segments backto-back If segment is lost, there will likely be many duplicate ACKs. If sender receives 4 ACKs for the same data (3 dups), it supposes that segment after ACKed data was lost: fast retransmit: resend segment before timer expires When resent packet is ACKed before a timeout, go to Fast Recovery Mode: - Halve "CongWin" - Increase CongWin by 1 MSS per CongWin bytes sent and ACKed. Transport Layer 3-74 Fast retransmit algorithm: event: ACK received, with ACK field value of y if (y > SendBase) { SendBase = y if (there are currently not-yet-acknowledged segments) start timer } else { increment count of dup ACKs received for y if (count of dup ACKs received for y = 3) { resend segment with sequence number y then resume sending new segments CongWin = CongWin/2 } a duplicate ACK for fast retransmit already ACKed segment Transport Layer 3-75 Fast Retransmit (after 3 duplicate ACKs) ( not CongWin limited) Sequence Number (kBytes) Retransmission Retransmission Retransmission. Send time (ms) Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-77 TCP Flow Control receive side of TCP connection has a receive buffer: flow control sender won’t overflow receiver’s buffer by transmitting too fast <- byte no. speed-matching higher byte no.s app process may be slow at reading from buffer service: matching the send rate to the receiving app’s drain rate Every TCP header sent by the receiver contains the value of the receiver's Window. This decreases as the incoming data buffer fills up. Transport Layer 3-78 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-79 TCP Connection Management Recall: TCP sender, receiver Three way handshake: establish “connection” before exchanging data segments initialize TCP variables: seq. #s buffers, flow control info (e.g. RcvWindow) client: connection initiator Socket clientSocket = new Socket("hostname","port number"); server: contacted by client Socket connectionSocket = welcomeSocket.accept(); Step 1: client host sends TCP SYN segment to server specifies initial seq # TCP option sends client’s MSS* no data Step 2: server host receives SYN, replies with SYN-ACK segment server allocates buffers specifies server initial seq. # TCP option sends server’s MSS Step 3: client receives SYN-ACK, replies with ACK segment, normally has no data MSS - maximum segment size, usually 1420 bytes. Transport Layer 3-80 TCP Connection Management (cont.) Closing a connection: client closes socket: clientSocket.close(); client server close Step 1: client end system close Step 2: server receives FIN, replies with ACK. Closes connection, sends FIN. timed wait sends TCP FIN control segment to server closed Or, either host sends a packet with the RES (reset) flag bit set. This is frequent between Web servers and browsers, Transport Layer 3-81 TCP Connection Management (cont) TCP server lifecycle TCP client lifecycle Transport Layer 3-82 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-83 Principles of Congestion Control Congestion: informally: “too many sources sending too much data too fast for network to handle” different from flow control! manifestations: lost packets (buffer overflow at routers) • Sender's CongWin decreases , throughput = CongWin/RTT decreases long delays (queueing in router buffers) • RTT increases, throughput = Window/RTT decreases Transport Layer 3-84 Causes/costs of congestion: scenario 1 Host A two senders, two receivers one router, infinite buffers no retransmission Host B lout lin : original data unlimited shared output link buffers large delays when congested maximum achievable throughput Transport Layer 3-85 Causes/costs of congestion: scenario 3 H o s t A l o u t H o s t B Another “cost” of congestion: when packet dropped, any “upstream transmission capacity used for that packet (to the point where it was dropped) was wasted! Transport Layer 3-86 Approaches towards congestion control Two broad approaches towards congestion control: End-end congestion control: no explicit feedback from network congestion inferred from end-system (1) observed loss, and (2) delay – by two methods: (a) duplicate ACKs or (b) RTOs. approach taken by TCP Network-assisted congestion control: routers provide feedback to end systems single bit indicating congestion (SNA, DECbit, TCP/IP ECN, ATM) explicit rate sender should send at Transport Layer 3-87 Case study: ATM ABR congestion control ABR: available bit rate: “elastic service” RM (resource management) cells: if sender’s path sent by sender, interspersed “underloaded”: sender should use available bandwidth if sender’s path congested: sender throttled to minimum guaranteed rate with data cells bits in RM cell set by switches (“network-assisted”) NI bit: no increase in rate (mild congestion) CI bit: congestion indication RM cells returned to sender by receiver, with bits intact Transport Layer 3-88 Chapter 3 outline 3.1 Transport-layer services 3.2 Multiplexing and demultiplexing 3.3 Connectionless transport: UDP 3.4 Principles of reliable data transfer 3.5 Connection-oriented transport: TCP segment structure reliable data transfer flow control connection management 3.6 Principles of congestion control 3.7 TCP congestion control Transport Layer 3-89 TCP congestion control: additive increase, multiplicative decrease Approach: increase transmission rate (window size), probing for usable bandwidth, until loss occurs additive increase: increase CongWin by 1 MSS every RTT until loss detected multiplicative decrease: cut CongWin in half after loss congestion window Saw tooth behavior: probing for bandwidth congestion window size 24 Kbytes Lost packet (3 duplicate ACKs) 16 Kbytes 8 Kbytes * MSS = maximum segment size (e.g.. , 1280 bytes). specified as a TCP option in SYN and SYN-ACK packets. time time Transport Layer 3-90 TCP Congestion Control: details sender limits transmission: LastByteSent-LastByteAcked CongWin Roughly, rate = CongWin Bytes/sec RTT CongWin is dynamic, function of perceived network congestion How does sender perceive congestion? loss event = timeout (RTO) or 3 duplicate ACKs TCP sender reduces rate (CongWin) after loss event mechanisms: AIMD (additive increase, multiplicative decrease) slow start (after RTO) conservative after timeout events (CongWin = 1 MSS) Transport Layer 3-91 TCP Slow Start When connection begins, CongWin = 1 MSS* Example: MSS = 500 bytes & RTT = 200 msec initial rate = 20 kbps available bandwidth may be >> MSS/RTT desirable to quickly ramp up to respectable rate When connection begins, increase rate exponentially fast until first loss event (doubles every RTT) or until CongWin = Receiver-Window. * MSS = maximum segment size (e.g.. , 1420 bytes). specified as a TCP option in SYN and SYN-ACK packets. Transport Layer 3-92 TCP Slow Start (more) When connection begins, Host B RTT increase rate exponentially until first loss event: Host A double CongWin every RTT done by incrementing CongWin for every ACK received (by d-SendBase) Summary: initial rate is slow but ramps up exponentially fast time Transport Layer 3-93 Q: When should the exponential increase switch to linear? A: When CongWin gets to 1/2 of its value before timeout. Implementation: Congestion Window /mss Refinement Variable Threshold “Fast Recovery” (after 3 dup. ACKs) 1/2 “Slow Start” Transmission Round (time/rtt) At loss event, Threshold is set to 1/2 of CongWin just before loss event Transport Layer 3-94 CongWin <= Threshold: Doubles each RTT (add MSS for each MSS ACKed) CongWin > Threshold: Adds MSS each RTT Time Out: Threshold = 1/2 CongWin, CongWin = 1 (Slow-Start) 3-Dup Ack: Threshold = 1/2 CongWin, CongWin = Threshold (Fast Recovery) CongWin / mss Time Out CongWin = 20 3 Dup. ACKs 12 6 Transport Layer 3-96 Refinement: inferring loss After 3 dup ACKs: is cut in half window then grows linearly But after timeout event: CongWin instead set to 1 MSS; window then grows exponentially to a threshold, then grows linearly CongWin Philosophy: 3 dup ACKs indicates network capable of delivering some segments timeout indicates a “more alarming” congestion scenario Transport Layer 3-96 Summary: TCP Congestion Control When CongWin is below Threshold, sender in slow- start phase, window grows exponentially. When CongWin is above Threshold, sender is in congestion-avoidance phase, window grows linearly. When a triple duplicate ACK occurs, Threshold set to CongWin/2 and CongWin set to Threshold (no slow start, linear mode immediately). When timeout occurs, Threshold set to CongWin/2 and CongWin is set to 1 MSS. Transport Layer 3-97 TCP sender congestion control State Event TCP Sender Action Commentary Slow Start (SS) ACK receipt for previously unacked data CongWin = CongWin + MSS, If (CongWin > Threshold) set state to “Congestion Avoidance” Resulting in a doubling of CongWin every RTT Congestion Avoidance (CA) ACK receipt for previously unacked data CongWin = CongWin+MSS * (MSS/CongWin) Additive increase, resulting in increase of CongWin by 1 MSS every RTT SS or CA Loss event detected by triple duplicate ACK Threshold = CongWin/2, CongWin = Threshold, Set state to “Congestion Avoidance” Fast recovery, implementing multiplicative decrease. CongWin will not drop below 1 MSS. SS or CA Timeout Threshold = CongWin/2, CongWin = 1 MSS, Set state to “Slow Start” Enter slow start SS or CA Duplicate ACK Increment duplicate ACK count for segment being ACKed CongWin and Threshold not changed Transport Layer 3-98 TCP throughput What’s the average throughout of TCP as a function of window size and RTT? Ignore slow start Let W be the window size when loss occurs. When window is W, throughput is W/RTT Just after loss, window drops to W/2, throughput to W/2RTT. Average throughout: .75 W/RTT (assuming loss occurs every time CongWin grows to W) Transport Layer 3-99 TCP Futures: TCP over “long, fat pipes” Example: With 1500 byte segments and 100ms RTT, if we want 10 Gbps throughput: Requires window size, W = 125,000,000 bytes (for 83,333 1500-byte in-flight segments) New versions of TCP for high-speed needed! Present W of 65,000 bytes -> 5 Mbps The TCP Window-Multiplier option allows the Window to be incearsed by a exponent of 2. Transport Layer 3-100 TCP Fairness Fairness goal: if K TCP sessions share same bottleneck link of bandwidth R, each should have average rate of R/K TCP connection 1 TCP connection 2 bottleneck router capacity R Transport Layer 3-101 Why is TCP fair? Two competing sessions: Additive increase gives slope of 1, as throughout increases multiplicative decrease decreases throughput proportionally Rmax Finally: equal bandwidth share loss: decrease windows by factor of 2 congestion avoidance: additive increase R1,R2 -losses: decrease windows by factor of 2 congestion avoidance: additive increase R1/2,R2/2 Start: R1 = 5 x R2, R1 +R2 < Rmax Start Connection 1 throughput Rmax Transport Layer 3-102 Fairness (more) Fairness and UDP Multimedia apps often do not use TCP do not want rate throttled by congestion control Instead use UDP: pump audio/video at constant rate, tolerate packet loss Research area: make UDP more TCP friendly Solution: reserve 50% of router buffer space for TCP segments (excess UDP segments dropped). Fairness and parallel TCP connections nothing prevents app from opening parallel connections between 2 hosts. Web browsers do this Example: link of rate R supporting 9 connections; new app asks for 1 TCP, gets rate R/10 new app asks for 10 TCPs, gets R/2 ! Browsers using parallel connections are not fair, but work better. Fairness is not an issue if there is no congestion. Transport Layer 3-103 Fixed congestion window (1) First case: WS > R * RTT (Rate Limited) ACK for first segment in window returns before window’s worth of data sent (so no idle time) Data sent at full rate, R R * RTT is the Bandwidth-Delay Product Transport Layer 3-104 Fixed congestion window (2) Second case: WS < R*RTT (Window Limited) wait for ACK after sending window’s worth of data sent Data sent at rate WS/RTT < R Transport Layer 3-105 TCP Delay Modeling: Slow Start (2) Delay components: • 2 RTT for connection estab and request • O/R to transmit object • time server idles due to slow start initiate TCP connection request object first window = S/R RTT second window = 2S/R Server idles: for RTT third window = 4S/R fourth window = 8S/R •O is object size, •R is transmission rate complete transmission object delivered •RTT is round-trip-time. time at client time at server Transport Layer 3-106 TCP Delay Modeling (3) As the CongWin grows, the connection can change from being "Window Limited" (DataRate=W/RTT) to being "Transmission Rate Limited" (DataRate = R). initiate TCP connection request object first window = S/R RTT second window = 2S/R third window = 4S/R fourth window = 8S/R complete transmission object delivered time at client time at server Transport Layer 3-107 32 bits TCP Options source port # Options appear in extra bytes appended to the end of the TCP header: number of bytes, option ID, parameter. A receiver can ignore Options it does not understand (extensible). dest port # sequence number acknowledgement number head not UA P R S F len used checksum Receive window Urg data pnter 20 Options (variable length) data <60 Common Options Time Stamp (RFC1323) Window Scale Option (RFC1323) Multiply Window by 2^n. SACK (RFC2018) Acknowledge non-contiguous blocks received. NOP - Pads out options to multiple of 4 bytes 3-108 Chapter 3: Summary principles behind transport layer services: multiplexing, demultiplexing reliable data transfer flow control congestion control instantiation and implementation in the Internet UDP TCP Next: leaving the network “edge” (application, transport layers) into the network “core” Transport Layer 3-109