Transport Layer 3-1 ECE 5650 TL: Reliable Data Transfer

Transport Layer 3-2 Recap: transport-layer protocols r IP service model m Host-to-host best-effort delivery service, datagram m Unreliable serivce r UDP: unreliable, unordered delivery m Proc to proc segment delivery: no- frills extension of “best-effort” IP (Multiplexing /demultiplexing) m Error checking r TCP: m reliable, in-order delivery: flow control, sequence number, ack, timer, etc m congestion control to prevent TCP connection from swamping the links and switches r services not available in either TCP or UDP: m delay guarantees m bandwidth guarantees application transport network data link physical 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 logical end-end transport

Transport Layer 3-3 Recap: Multiplexing/demultiplexing (encode/decode) application transport network link physical P1 application transport network link physical application transport network link physical P2 P3 P4 P1 host 1 host 2 host 3 = process= socket The job of delivering received transport-layer segments to correct socket. Each socket has a unique identifier (a set of fields in each segment) Demultiplexing at rcv host: The job of gathering data from different sockets, enveloping data with header (later used for demultiplexing) and passing the data to the network layer. Multiplexing at send host:

Transport Layer 3-4 Recap: Connectionless demux DatagramSocket serverSocket = new DatagramSocket(6428); Client IP:B P2 client IP: A P1 P3 server IP: C SP: 6428 DP: 9157 SP: 9157 DP: 6428 SP: 6428 DP: 5775 SP: 5775 DP: 6428 SP = Source Port, DP = Destination Port Source Packet provides “return address” that destination host can use to reply port 9157 port 6428port 5775 One Server Socket

Transport Layer 3-5 Recap: Conn-oriented demux Client IP:B P1 client IP: A P1P2P4 server IP: C SP: 9157 DP: 80 SP: 9157 DP: 80 P5P6P3 D-IP:C S-IP: A D-IP:C S-IP: B SP: 5775 DP: 80 D-IP:C S-IP: B SP = Source Port, DP = Destination Port, S-IP = Source IP, D-IP = Destination IP Three Server Sockets

Transport Layer 3-6 Recap: UDP Segment Structure r UDP segment has an 8 byte header. r Length of UDP segment, in bytes, including header. r Minimum length is 8 and maximum is 64k. source port #dest port # 32 bits Application data (message) UDP segment format length (16-bit) checksum (16-bit)

Transport Layer 3-7 Recap: Internet Checksum: Sender Example r Note m When adding numbers, a carryout from the most significant bit needs to be added to the result r Segment has two 16-bit integers and the checksum r Add the two 16-bit integers r compute the checksum by taking the 1’s complement of the sum wraparound: add carry to LSB sum checksum

Transport Layer 3-8 TL outline r 3.1 Transport-layer services r 3.2 Multiplexing and demultiplexing r 3.3 Connectionless transport: UDP r 3.4 Principles of reliable data transfer r 3.5 Connection-oriented transport: TCP m segment structure m reliable data transfer m flow control m connection management r 3.6 Principles of congestion control r 3.7 TCP congestion control

Transport Layer 3-9 Principles of Reliable data transfer r Reliable transfer: no data corrupted or loss, and in-order delivery m important in app., transport, link layers m top-10 list of important networking topics! r characteristics of unreliable channel will determine complexity of reliable data transfer protocol (rdt)

Transport Layer 3-10 Reliable data transfer: getting started send side receive side rdt_send(): called from above, (e.g., by app.). Passed data to deliver to receiver upper layer udt_send(): called by rdt, to transfer packet over unreliable channel to receiver rdt_rcv(): called when packet arrives on rcv-side of channel deliver_data(): called by rdt to deliver data to upper

Transport Layer 3-11 Reliable data transfer: getting started We’ll: r incrementally develop sender, receiver sides of reliable data transfer protocol (rdt) r consider only unidirectional data transfer m but control info will flow on both directions! r use finite state machines (FSM) to specify sender, receiver state 1 state 2 event causing state transition (when..happens) actions taken on state transition (then this happens) state: when in this “state” next state uniquely determined by next event event actions Symbol  above or below the line denote the lack of an action or event

Transport Layer 3-12 Finite State Machine r FSM is a model of behavior composed of states, transitions and events m State stores info about the past, reflecting the system start to the present moment m Transition indicates a state change, triggered by an event m Events: activity to be performed at a given moment: entry, exit, input r FSM often represented using a state diagram

Transport Layer 3-13 Rdt1.0: reliable transfer over a reliable channel r underlying channel perfectly reliable m no bit errors m no loss of packets r separate FSMs for sender, receiver: m sender sends data into underlying channel m receiver read data from underlying channel Wait for call from above packet = make_pkt(data) udt_send(packet) rdt_send(data) extract (packet,data) deliver_data(data) Wait for call from below rdt_rcv(packet) sender receiver

Transport Layer 3-14 Rdt2.0: channel with bit errors but does not lose packets r underlying channel may flip bits in packet m Solution: checksum to detect bit errors r the question: how to recover from errors: m acknowledgements (ACKs): receiver explicitly tells sender that pkt received “OK” m negative acknowledgements (NAKs): receiver explicitly tells sender that pkt had errors “please repeat that”. m sender retransmits pkt on receipt of NAK  new mechanisms in rdt2.0 (beyond rdt1.0 ): m error detection m receiver feedback: control msgs (ACK,NAK) rcvr->sender

Transport Layer 3-15 Automatic Repeat reQuest protocols (ARQ) r Protocols based on ACKs and NAKs. r ARQ Protocols have 3 capabilities: m Error detection of bit errors which requires additional bits added to the transmitted packet. m Receiver feedback: ACK and NAK bit (1=ACK, 0=NAK) to be added to the transmitted packet. m Retransmission: a packet with errors will be retrasmitted.

Transport Layer 3-16 rdt2.0: FSM specification Wait for call from above sndpkt = make_pkt(data, checksum) udt_send(sndpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) rdt_rcv(rcvpkt) && isACK(rcvpkt) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) udt_send(NAK) rdt_rcv(rcvpkt) && corrupt(rcvpkt) Wait for ACK or NAK Wait for call from below sender receiver rdt_send(data)  Sender can not send another packet until receiver sends back ACK/NAK. (send-and-wait protocols)

Transport Layer 3-17 rdt2.0: operation with no errors Wait for call from above snkpkt = make_pkt(data, checksum) udt_send(sndpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) rdt_rcv(rcvpkt) && isACK(rcvpkt) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) udt_send(NAK) rdt_rcv(rcvpkt) && corrupt(rcvpkt) Wait for ACK or NAK Wait for call from below rdt_send(data) 

Transport Layer 3-18 rdt2.0: error scenario Wait for call from above snkpkt = make_pkt(data, checksum) udt_send(sndpkt) extract(rcvpkt,data) deliver_data(data) udt_send(ACK) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) rdt_rcv(rcvpkt) && isACK(rcvpkt) udt_send(sndpkt) rdt_rcv(rcvpkt) && isNAK(rcvpkt) udt_send(NAK) rdt_rcv(rcvpkt) && corrupt(rcvpkt) Wait for ACK or NAK Wait for call from below rdt_send(data) 

Transport Layer 3-19 rdt2.0 has a fatal flaw! What happens if ACK/NAK corrupted? r sender doesn’t know what happened at receiver! r can’t just retransmit: possible duplicate Handling duplicates: r sender retransmits current packet if ACK/NAK garbled/corrupt r sender adds sequence number to each packet, 1-bit that changes if a new packet is sent r receiver discards (doesn’t deliver up) duplicate pkt r ACK/NAK packets do not need the sequence number of the packet acknowledged since sender knows it is for the last packet (we assume no packet loss)

Transport Layer 3-20 rdt2.1: sender, handles garbled ACK/NAKs Wait for call 0 from above sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) rdt_send(data) Wait for ACK or NAK 0 udt_send(sndpkt) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isNAK(rcvpkt) ) sndpkt = make_pkt(1, data, checksum) udt_send(sndpkt) rdt_send(data) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt) udt_send(sndpkt) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isNAK(rcvpkt) ) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt) Wait for call 1 from above Wait for ACK or NAK 1  

Transport Layer 3-21 rdt2.1: receiver, handles garbled ACK/NAKs Wait for 0 from below sndpkt = make_pkt(NAK, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && not corrupt(rcvpkt) && has_seq0(rcvpkt) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq1(rcvpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) Wait for 1 from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq0(rcvpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && not corrupt(rcvpkt) && has_seq1(rcvpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt)) sndpkt = make_pkt(ACK, chksum) udt_send(sndpkt) sndpkt = make_pkt(NAK, chksum) udt_send(sndpkt)

Transport Layer 3-22 rdt2.1: discussion Sender: r seq # added to pkt r two seq. #’s (0,1) will suffice. Why? r must check if received ACK/NAK corrupted r twice as many states m state must “remember” whether “current” pkt has 0 or 1 seq. # Receiver: r must check if received packet is duplicate m state indicates whether 0 or 1 is expected pkt seq # r note: receiver can not know if its last ACK/NAK received OK at sender

Transport Layer 3-23 rdt2.2: a NAK-free protocol r same functionality as rdt2.1, using ACKs only r instead of NAK, receiver sends ACK for last pkt received OK m receiver must explicitly include seq # of pkt being ACKed r duplicate ACK at sender results in same action as NAK: retransmit current pkt

Transport Layer 3-24 rdt2.2: sender, receiver fragments Wait for call 0 from above sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) rdt_send(data) udt_send(sndpkt) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isACK(rcvpkt,1) ) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,0) Wait for ACK 0 sender FSM fragment Wait for 0 from below rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq1(rcvpkt) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(ACK1, chksum) udt_send(sndpkt) rdt_rcv(rcvpkt) && (corrupt(rcvpkt) || has_seq1(rcvpkt)) udt_send(sndpkt) receiver FSM fragment 

Transport Layer 3-25 rdt3.0: channels with errors and loss New assumption: underlying channel can also lose packets (data or ACKs) m checksum, seq. #, ACKs, retransmissions will be of help, but not enough Approach: sender waits “reasonable” amount of time for ACK r retransmits if no ACK received in this time r if pkt (or ACK) just delayed (not lost): m retransmission will be duplicate, but use of seq. #’s already handles this m receiver must specify seq # of pkt being ACKed r requires countdown timer timeout is worst-case time (taking all types of delays into account) for packet to reach receiver and for an ACK to be received

Transport Layer 3-26 rdt3.0 sender sndpkt = make_pkt(0, data, checksum) udt_send(sndpkt) start_timer rdt_send(data) Wait for ACK0 rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isACK(rcvpkt,1) ) Wait for call 1 from above sndpkt = make_pkt(1, data, checksum) udt_send(sndpkt) start_timer rdt_send(data) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,0) rdt_rcv(rcvpkt) && ( corrupt(rcvpkt) || isACK(rcvpkt,0) ) rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && isACK(rcvpkt,1) stop_timer udt_send(sndpkt) start_timer timeout udt_send(sndpkt) start_timer timeout rdt_rcv(rcvpkt) Wait for call 0from above Wait for ACK1  rdt_rcv(rcvpkt)   

Transport Layer 3-27 rdt3.0 in action Alternating-bit protocol because packet sequence numbers alternate between 0 and 1

Transport Layer 3-28 rdt3.0 in action

Transport Layer 3-29 Performance of rdt3.0 r rdt3.0 works, but performance stinks r example: 1 Gbps link, 15 ms e-e propagation delay, 1KByte packet: T transmit = 8kb/pkt 10^9 b/sec = 8 microsec m U sender : utilization – fraction of time sender busy sending m 1 KByte packet every 30 msec -> 267 kbps throughput over 1 Gbps link m network protocol limits use of physical resources! L (packet length in bits) R (transmission rate, bps) = last bit arrives at receiver at 15 ms + 8 microsec = ms ACK comes back at sender at ms + 15 ms = ms (assuming ACK size is small)

Transport Layer 3-30 rdt3.0: stop-and-wait operation first packet bit transmitted, t = 0 senderreceiver RTT last packet bit transmitted, t = L / R first packet bit arrives last packet bit arrives, send ACK ACK arrives, send next packet, t = RTT + L / R

Transport Layer 3-31 Pipelined protocols Pipelining: sender allows multiple, “in-flight”, yet-to- be-acknowledged pkts m range of sequence numbers must be increased m buffering at sender and/or receiver r Two generic forms of pipelined protocols: go-Back-N, selective repeat

Transport Layer 3-32 Pipelining: increased utilization first packet bit transmitted, t = 0 senderreceiver RTT last bit transmitted, t = L / R first packet bit arrives last packet bit arrives, send ACK ACK arrives, send next packet, t = RTT + L / R last bit of 2 nd packet arrives, send ACK last bit of 3 rd packet arrives, send ACK Increase utilization by a factor of 3!

Transport Layer 3-33 Go-Back-N (Sliding Window) Sender: r There is a k-bit sequence # in packet header r “window” of up to N, consecutive unacknowledged sent/can-be-sent packets allowed r window moves by 1 packet at a time when its 1st sent pkt is acknowledged (standard behavior) Sender must respond to three types of events: r 1- Invocation from above: application layers tries to send a packet, if window is full then packet is returned otherwise the packet is accepted and sent. r 2- Receipt of an ACK: One ACK(n) received indicates that all pkts up to, including seq # n have been received - “cumulative ACK” m may receive duplicate ACKs (when receiver receives out-of-order packets) r 3- A timeout event (only cause of retransmission): m timer for each in-flight pkt. m if timeout occurs: retransmit packets that have not been acknowledged. window cannot contain acknowledged pkts

Transport Layer 3-34 GBN: sender extended FSM Wait start_timer udt_send(sndpkt[base]) udt_send(sndpkt[base+1]) … udt_send(sndpkt[nextseqnum-1]) timeout rdt_send(data) if (nextseqnum < base+N) { sndpkt[nextseqnum] = make_pkt(nextseqnum,data,chksum) udt_send(sndpkt[nextseqnum]) if (base == nextseqnum) start_timer nextseqnum++ } else refuse_data(data) base = getacknum(rcvpkt)+1 If (base == nextseqnum) stop_timer else start_timer rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) base=1 nextseqnum=1 rdt_rcv(rcvpkt) && corrupt(rcvpkt) 

Transport Layer 3-35 GBN: receiver extended FSM ACK-only: always send ACK for correctly-received pkt with highest in-order seq #  need only remember expectedseqnum r When out-of-order pkt is received: m discard (don’t buffer) -> no receiver buffering! m Re-ACK pkt with highest in-order sequence # Wait udt_send(sndpkt) default rdt_rcv(rcvpkt) && notcurrupt(rcvpkt) && hasseqnum(rcvpkt,expectedseqnum) extract(rcvpkt,data) deliver_data(data) sndpkt = make_pkt(expectedseqnum,ACK,chksum) udt_send(sndpkt) expectedseqnum++ expectedseqnum=1 sndpkt = make_pkt(expectedseqnum,ACK,chksum) 

Transport Layer 3-36 GBN in action Window-size = 4

Transport Layer 3-37 Notes on GBN: r Receiver discards out-of-order packets and does not even buffer them for future use: m Receiver must deliver data in order to application layer m If out-of-order packets are buffered and in-order packets are lost then a retransmission will re-send the out-of-order packets. m Simplicity of receiver buffering due to discarding m One disadvantage of not buffering is that a retransmission could be also lost which would require additional retransmissions. r A single packet error can cause GBN sender to retransmit larger #pkts as the window size increases.

Transport Layer 3-38 Selective Repeat r receiver individually acknowledges all correctly received pkts m buffers pkts, as needed, for eventual in-order delivery to upper layer r sender only resends pkts for which ACK not received m sender timer for each unACKed pkt r sender window m N consecutive seq #’s m again limits seq #s of sent, unACKed pkts

Transport Layer 3-39 Selective repeat: sender, receiver windows Window may contain acknowledged pkts (unlike GBN)

Transport Layer 3-40 Selective repeat data from above : r if next available seq # in window, send pkt timeout(n): r resend pkt n, restart timer ACK(n) in window: r mark pkt n as received r if n smallest unACKed pkt, advance window base to next unACKed seq # pkt n in window r send ACK(n) r out-of-order: buffer r in-order: deliver (also deliver buffered, in-order pkts), advance window to next not-yet-received pkt pkt n previously acknowledged r ACK(n), necessary to move sender’s window otherwise: r ignore receiver eventssender events

Transport Layer 3-41 Selective repeat in action

Transport Layer 3-42 Selective repeat: too large window dilemma Example: r seq #’s: 0, 1, 2, 3 r window size=3 r receiver sees no difference in two scenarios! r incorrectly passes duplicate data as new in (a) r To avoid this condition:the sequence number space (k) must be at least twice as large as the window size (w) or k >= 2w

Transport Layer 3-43 Reliable Data Transfer Mechanisms r Checksum: used to detect bit errors in transmitted pkt. r Timer: used to retransmit a packet when the packet or its ACK was lost. Duplicate copies of pkt could be received due to premature timeouts when the packet or its ACK are delayed. r Sequence number: used for sequential numbering of packets sent. Gaps in seq #s indicate lost/delayed packts while duplicate seq #s indicate duplicate pkts. r Acknowledgement (ACK): used by receiver to tell sender a packet (individual ACK) or packets (cumulative) have been received. r Negative Acknowledgement (NAK): used by receiver to tell sender a packet has not been received correctly. r Pipelining: used to allow sending packets even if prior ones were not acknowledged and hence increasing the utilization of the sender. r Window: used to restrict sender to send packets with seq #s within a range (window size) without waiting for an acknowledgement.

Transport Layer 3-44 Summary r Automatic Repeat reQuest protocols (ARQ) r Send-and-wait protocols r The use of Sequences numbers in reliable data transfers r Low Utilization of the Send-and-wait protocols r 2 approaches to Pipelined error recovery: m Go-Back-N (GBN)/sliding-window protocol m Selective Repeat (SR)