Chapter 3 Transport Layer

Transcription

1 Chapter 3 Transport Layer A note on the use of these ppt slides: We re making these slides freely aailable to all (faculty, students, readers). They re in PowerPoint form so you see the animations; and can add, modify, and delete slides (including this one) and slide content to suit your needs. They obiously 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) that you mention their source (after all, we d like people to use our book!) If you post any slides 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. Thanks and enjoy! JFK/KWR All material copyright J.F Kurose and K.W. Ross, All Rights Resered Computer Networking: A Top Down Approach 6 th edition Jim Kurose, Keith Ross Addison-Wesley April 2012 Transport Layer 3-1

2 Chapter 3: Transport Layer learn about Internet transport layer protocols: UDP: connectionless transport, brief not much to say TCP: connection-oriented reliable transport TCP congestion control Transport Layer 3-2

3 Transport serices and protocols proide 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 rc side: reassembles segments into messages, passes to app layer more than one transport protocol aailable to apps Internet: TCP and UDP no delay/serice guarantees application transport network data link physical logical end-end transport application transport network data link physical Transport Layer 3-3

4 Multiplexing/demultiplexing multiplexing at sender: handle data from multiple sockets, add transport header (later used for demultiplexing) demultiplexing at receier: use header info to delier receied segments to correct socket application application P3 transport network link P1 P2 transport network link physical application P4 transport network link socket process physical physical Transport Layer 3-4

5 UDP: User Datagram Protocol [RFC 768] no frills, bare bones Internet transport protocol best effort serice, UDP segments may be: lost deliered out-of-order to app connectionless: no handshaking between UDP sender, receier each UDP segment handled independently of others UDP use: streaming multimedia apps (loss tolerant, rate sensitie) DNS SNMP reliable transfer oer UDP: add reliability at application layer application-specific error recoery! Transport Layer 3-5

6 UDP: segment header 32 bits source port # dest port # length application data (payload) checksum UDP segment format length, in bytes of UDP segment, including header why is there a UDP? no connection establishment (which can add delay) simple: no connection state at sender, receier small header size no congestion control: UDP can blast away as fast as desired Transport Layer 3-6

7 Transport Layer 3-7

8 Connection-oriented demux TCP socket identified by 4-tuple: source IP address source port number dest IP address dest port number demux: receier uses all four alues to direct segment to appropriate socket serer host may support many simultaneous TCP sockets: each socket identified by its own 4-tuple web serers hae different sockets for each connecting client non-persistent HTTP will hae different socket for each request Transport Layer 3-8

9 Connection-oriented demux threaded serer application application P3 transport network link physical P4 transport network link physical serer: IP address B application P2 P3 transport network link physical host: IP address A source IP,port: B,80 dest IP,port: A,9157 source IP,port: C,5775 dest IP,port: B,80 host: IP address C source IP,port: A,9157 dest IP, port: B,80 source IP,port: B,9157 dest IP,port: B,80 Transport Layer 3-9

10 TCP segment structure URG: urgent data (generally not used) ACK: ACK # alid PSH: push data now (generally not used) RST, SYN, FIN: connection estab (setup, teardown commands) Internet checksum (as in UDP) 32 bits source port # dest port # head len sequence number acknowledgement number not used UAP R S F checksum receie window application data (ariable length) Urg data pointer options (ariable length) counting by bytes of data (not segments!) # bytes rcr willing to accept Transport Layer 3-10

11 TCP: Oeriew RFCs: 793,1122,1323, 2018, 2581 point-to-point: one sender, one receier reliable, in-order byte steam: no message boundaries pipelined: TCP congestion and flow control set window size Hybrid of go back n and selectie repeat full duplex data: bi-directional data flow in same connection MSS: maximum segment size connection-oriented: handshaking (exchange of control msgs) inits sender, receier state before data exchange flow controlled: sender will not oerwhelm receier Transport Layer 3-11

12 channels with errors and loss assumption: underlying channel can lose packets (data, 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 receied in this time if pkt (or ACK) just delayed (not lost): retransmission will be duplicate, but seq. # s already handles this receier must specify seq # of pkt being ACKed requires countdown timer Transport Layer 3-12

13 Pipelined protocols pipelining: sender allows multiple, in-flight, yetto-be-acknowledged pkts range of sequence numbers must be increased buffering at sender and/or receier two generic forms of pipelined protocols: go-back-n, selectie repeat Transport Layer 3-13

14 Pipelined protocols: concepts Go-back-N: sender can hae up to N unacked packets in pipeline receier only sends cumulatie ack doesn t ack packet if there s a gap sender has timer for oldest unacked packet when timer expires, retransmit all unacked packets Selectie Repeat: sender can hae up to N unack ed packets in pipeline rcr sends indiidual ack for each packet sender maintains timer for each unacked packet when timer expires, retransmit only that unacked packet Transport Layer 3-14

15 TCP seq. numbers, ACKs sequence numbers: byte stream number of first byte in segment s data acknowledgements: seq # of next byte expected from other side cumulatie ACK Q: how receier handles out-of-order segments A: TCP spec doesn t say, outgoing segment from sender source port # dest port # sequence number acknowledgement number rwnd checksum sent ACKed urg pointer window size N sender sequence number space sent, notyet usable not ACKed but not usable ( inflight ) yet sent incoming segment to sender - up to implementor source port # dest port # sequence number acknowledgement number A rwnd checksum urg pointer Transport Layer 3-15

16 TCP seq. numbers, ACKs Host A Host B User types C host ACKs receipt of echoed C Seq=42, ACK=79, data = C Seq=79, ACK=43, data = C Seq=43, ACK=80 host ACKs receipt of C, echoes back C simple telnet scenario Transport Layer 3-16

17 Transport Layer 3-17

18 TCP round trip time, timeout Q: how to set TCP timeout alue? longer than RTT but RTT aries too short: premature timeout, unnecessary retransmissions too long: slow reaction to segment loss Q: how to estimate RTT? SampleRTT: measured time from segment transmission until ACK receipt ignore retransmissions SampleRTT will ary, want estimated RTT smoother aerage seeral recent measurements, not just current SampleRTT Transport Layer 3-18

19 TCP round trip time, timeout EstimatedRTT = (1- a)*estimatedrtt + a*samplertt exponential weighted moing aerage influence of past sample decreases exponentially fast RTT: gaia.cs.umass.edu to fantasia.eurecom.fr typical alue: a = RTT: gaia.cs.umass.edu to fantasia.eurecom.fr RTT (milliseconds) RTT (milliseconds) samplertt EstimatedRTT time (seconnds) time (seconds) SampleRTT Estimated RTT Transport Layer 3-19

20 TCP round trip time, timeout timeout interal: EstimatedRTT plus safety margin large ariation in EstimatedRTT -> larger safety margin estimate SampleRTT deiation from EstimatedRTT: DeRTT = (1-b)*DeRTT + b* SampleRTT-EstimatedRTT (typically, b = 0.25) TimeoutInteral = EstimatedRTT + 4*DeRTT estimated RTT safety margin Transport Layer 3-20

21 TCP simple sender (no optimizations, congestion control): data rcd from app: create segment with seq # seq # is byte-stream number of first data byte in segment start timer if not already running think of timer as for oldest unacked segment expiration interal: TimeOutInteral timeout: retransmit segment that caused timeout restart timer ack rcd: if ack acknowledges preiously unacked segments update what is known to be ACKed start timer if there are still unacked segments Transport Layer 3-21

22 TCP: retransmission scenarios Host A Host B Host A Host B Host A Host B SendBase=92 Seq=92, 8 bytes of data Seq=92, 8 bytes of data Seq=92, 8 bytes of data timeout X ACK=100 timeout Seq=100, 20 bytes of data ACK=100 ACK=120 timeout Seq=100, 20 bytes of data ACK=100 X ACK=120 Seq=92, 8 bytes of data ACK=100 SendBase=100 SendBase=120 Seq=92, 8 bytes of data ACK=120 Seq=120, 15 bytes of data SendBase=120 lost ACK scenario premature timeout cumulatie ACK Transport Layer 3-22

23 TCP ACK original generation [RFC 1122, RFC 2581] eent at receier arrial of in-order segment with expected seq #. All data up to expected seq # already ACKed arrial of in-order segment with expected seq #. One other segment has ACK pending arrial of out-of-order segment higher-than-expect seq. #. Gap detected arrial of segment that partially or completely fills gap TCP receier action delayed ACK. Wait up to 500ms for next segment. If no next segment, send ACK immediately send single cumulatie ACK, ACKing both in-order segments immediately send duplicate ACK, indicating seq. # of next expected byte immediate send ACK, proided that segment starts at lower end of gap Transport Layer 3-23

24 TCP fast retransmit detect loss before timeout time-out period often relatiely long: long delay before resending lost packet detect lost segments ia duplicate ACKs. sender often sends many segments backto-back if segment is lost, there will likely be many duplicate ACKs. TCP fast retransmit if sender receies 3 ACKs for same data ( triple duplicate ACKs ), resend unacked segment with smallest seq # likely that unacked segment lost, so don t wait for timeout Transport Layer 3-24

25 TCP fast retransmit Host A Host B Seq=92, 8 bytes of data Seq=100, 20 bytes of data X timeout ACK=100 ACK=100 ACK=100 ACK=100 Seq=100, 20 bytes of data fast retransmit after sender receipt of triple duplicate ACK Transport Layer 3-25

26 Connection Management before exchanging data, sender/receier handshake : agree to establish connection (each knowing the other willing to establish connection) agree on connection parameters sequence number application connection state: ESTAB connection ariables: seq # client-to-serer serer-to-client rcbuffer size at serer,client network application connection state: ESTAB connection Variables: seq # client-to-serer serer-to-client rcbuffer size at serer,client network Socket clientsocket = newsocket("hostname","port number"); Socket connectionsocket = welcomesocket.accept(); Transport Layer 3-29

27 Agreeing to establish a connection 2-way handshake failure scenarios: choose x retransmit req_conn(x) req_conn(x) acc_conn(x) ESTAB choose x retransmit req_conn(x) req_conn(x) acc_conn(x) ESTAB ESTAB client terminates req_conn(x) connection x completes serer forgets x ESTAB retransmit data(x+1) client terminates data(x+1) connection x completes req_conn(x) accept data(x+1) serer forgets x half open connection! (no client!) ESTAB data(x+1) ESTAB accept data(x+1) Transport Layer 3-30

28 TCP 3-way handshake client state LISTEN SYNSENT ESTAB choose init seq num, x send TCP SYN msg receied SYNACK(x) indicates serer is lie; send ACK for SYNACK; this segment may contain client-to-serer data SYNbit=1, Seq=x SYNbit=1, Seq=y ACKbit=1; ACKnum=x+1 ACKbit=1, ACKnum=y+1 choose init seq num, y send TCP SYNACK msg, acking SYN receied ACK(y) indicates client is lie serer state LISTEN SYN RCVD ESTAB Transport Layer 3-31

29 Transport Layer 3-34

30 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 oerflow at routers) long delays (queueing in router buffers) a top-10 problem! Transport Layer 3-35

31 Causes/costs of congestion: delay two senders, two receiers one router, infinite buffers output link capacity: R no retransmission original data: l in Host A unlimited shared output link buffers throughput: l out Host B R/2 l out delay l in R/2 l in R/2 maximum per-connection throughput: R/2 large delays as arrial rate, l in, approaches capacity Transport Layer 3-36

32 Causes/costs of congestion: drop + duplicate one router, finite buffers sender retransmission of timed-out packet application-layer input = application-layer output: l in = l out transport-layer input includes retransmissions : l in l in l in : original data l' in : original data, plus retransmitted data l out Host A Host B finite shared output link buffers Transport Layer 3-37

33 Causes/costs of congestion: duplicates packets can be lost, dropped at router due to full buffers sender times out prematurely, sending two copies, both of which are deliered R/2 lout l in R/2 when sending at R/2, some packets are retransmissions including duplicated that are deliered! costs of congestion: more work (retrans) for gien goodput unneeded retransmissions: link carries multiple copies of pkt decreasing goodput Transport Layer 3-38

34 Causes/costs of congestion: multi-hop paths C/2 l out l in C/2 another cost of congestion: when packet dropped, any upstream transmission capacity used for that packet was wasted! Transport Layer 3-39

35 Approaches towards congestion control two broad approaches towards congestion control: end-end congestion control: no explicit feedback from network congestion inferred from end-system obsered loss, delay, Ack signals from receier approach taken by TCP network-assisted congestion control: routers proide feedback to end systems single bit indicating congestion as packet forwarded (SNA, DECbit, TCP/IP ECN) Transport Layer 3-40

36 TCP congestion control basics: additie increase multiplicatie decrease approach: sender increases transmission rate (window size), probing for usable bandwidth, until loss occurs additie increase: increase cwnd by 1 MSS (max segment size, default 536 bytes) eery RTT until loss detected multiplicatie decrease: cut cwnd in half after loss AIMD saw tooth behaior: probing for bandwidth cwnd: TCP sender congestion window size additiely increase window size. until loss occurs (then cut window in half) time Transport Layer 3-41

37 TCP Congestion Control: window sender sequence number space cwnd last byte ACKed sent, notyet ACKed ( in-flight ) last byte sent sender limits transmission: LastByteSent- LastByteAcked < cwnd TCP sending rate: roughly: send cwnd bytes, wait RTT for ACKS, then send more bytes rate ~ cwnd RTT bytes/sec cwnd is dynamic, function of perceied network congestion Transport Layer 3-42

38 TCP Slow Start when connection begins, rate (window) small but increase rate (window) exponentially until first loss eent: initially cwnd = 1 MSS double cwnd eery RTT done by incrementing cwnd for eery ACK receied summary: initial rate is slow but ramps up exponentially fast Host A RTT one segment two segments four segments Host B time Transport Layer 3-43

39 TCP: detecting and responding to loss If loss detected by timeout: cwnd set to 1 MSS window initially grows exponentially to threshold (ssthresh), then grows linearly Fast Retransmit (reminder): Faster loss detection signaled by duplicate ACKs from receier; Receier adjusted to send Ack each time packet receied een if it s a duplicate (no adance in seq number) Sender waits for 3 duplicate Acks (not just one in case slightly out of order deliery) Transport Layer 3-44

40 TCP: reacting to loss If loss detected by timeout cwnd set to 1 MSS If lost detected by duplicate ACKS: can be less reactie since ACKs indicate network able to delier packets to receier TCP Reno: cwnd is only cut in half (not set to 1); then window grows linearly TCP New Reno: improes retransmit during fast recoery gien wireless link loss is not usually due to network congestion for eery ACK that adances sequence space, send next packet beyond the ACKed sequence number as well Transport Layer 3-45

41 TCP: switching from slow start to CA Q: when should the exponential increase switch to linear? A: when cwnd gets to 1/2 of its alue before timeout. Implementation: ariable ssthresh on loss eent, ssthresh is set to 1/2 of cwnd just before loss eent Transport Layer 3-46

42 TCP throughput ag. TCP thruput as function of window size, RTT? ignore transient of slow start, assume always data to send W: window size (measured in bytes) where loss occurs ag. window size (# in-flight bytes) is ¾ W ag. thruput is 3/4W per RTT ag TCP thruput = 3 4 W RTT bytes/sec W W/2 Transport Layer 3-48

43 TCP Fairness fairness goal: if K TCP sessions share same bottleneck link of bandwidth R, each should hae aerage rate of R/K TCP connection 1 TCP connection 2 bottleneck router capacity R Transport Layer 3-49

44 Why is TCP fair, in concept? two competing sessions: additie increase gies slope of 1, as throughout increases multiplicatie decrease decreases throughput proportionally R equal bandwidth share Connection 2 throughput Connection 1 throughput loss: decrease window by factor of 2 congestion aoidance: additie increase loss: decrease window by factor of 2 congestion aoidance: additie increase R Transport Layer 3-50