1 TCP. 2 Referencias Cerf, V., and R. Kahn, "A Protocol for Packet Network Intercommunication", IEEE Transactions on Communications, Vol. COM-22, No.

Presentación del tema: "1 TCP. 2 Referencias Cerf, V., and R. Kahn, "A Protocol for Packet Network Intercommunication", IEEE Transactions on Communications, Vol. COM-22, No."— Transcripción de la presentación:

1 1 TCP

2 2 Referencias Cerf, V., and R. Kahn, "A Protocol for Packet Network Intercommunication", IEEE Transactions on Communications, Vol. COM-22, No. 5, pp 637-648, May 1974. J. Postel, RFC-793 "Transmission Control Protocol" September 1981. ( Various errors and inconsistencies were detected) Clarifications and bug fixes in RFC 1122 October 1989 Extensions in RFC 1323 Peterson and Davie, pp 378-405

3 3 Agenda o Revisión Nivel de Transporte o El Protocolo TCP  Características  TCP Connection setup  Segmentos TCP  Números de Secuencia TCP  TCP Sliding Window  Control de Flujo  Timeouts y Retransmisiones (RTX)

4 4 Recordar los modelos de capas…. 1 Parte de la clase

5 5 Nivel Transporte EnriLean milagros.dc.uba.ar stone.ac.upc.es Nivel de Red Nivel de Enlace Nivel de Aplicación Nivel de Transporte O.S. HeaderDataHeaderData HD HD HD HDHD HD

6 6 Modelo OSI Session Network Link Physical Application Presentation Transport Network Link Network Transport Session Presentation Application Network Link Physical Peer-layer communication layer-to-layer communication Router 1 2 3 4 5 6 7 1 2 3 4 5 6 7

7 7 TCP : Características  TCP es orientado a conexion  Manejo de la conexión : 3-way handshake usado para setup y 2-2 o 4 way handshake para la liberación  TCP provee un servicio de flujo de bytes (stream-of-bytes )  TCP es confiable ( estableciendo una suerte de “conexión lógica entre los sockets”)  Acknowledgements ACKs  Checksums  Números de secuencia para detectar datos perdidos o desordenados  Datos perdidos o corruptos se RTX después de un timeout.  Datos desordenados se podrían reordenar.  Control de Flujo evita inundar al receptor.  TCP implementa mecanismos de control de congestión ( se le dedica una clase la semana próxima ).

8 8 TCP : orientado a conexión 3-way handshake (Activo) Cliente (Pasivo) Server Syn Syn + Ack Ack 4-way handshake (Activo) Cliente (Pasivo) Server Fin (Data +) Ack Fin Ack Caso ideal, en el RFC 793 Se plantean varios escenarios

9 9 TCP soporta un “stream de bytes” Byte 0Byte 1 Byte 2Byte 3 Byte 0Byte 1Byte 2Byte 3 Host A Host B Byte 80

10 10 …dicho servicio se emula usando segmentos TCP Byte 0Byte 1 Byte 2Byte 3 Byte 0Byte 1Byte 2Byte 3 Host A Host B Byte 80 TCP Data Byte 80 Un segmento se envía cuando: 1.Segmento full (MSS bytes), 2.No esta “full”, pero sucede time out, 3.“Pushed” por la aplicación.

11 11 Formato del Segmento TCP IP Hdr IP Data TCP HdrTCP Data Src portDst port Sequence # Ack Sequence # HLEN 4 RSVD 6 URGACK PSH RSTSYNFIN Flags Window Size ChecksumUrg Pointer (TCP Options) 01531 TCP Data TCP Header y Data + dirección IP Los números de port Src/dst y la direcciones identifican unívocamente un “socket” Veremos concepto de pseudo header

12 12 Números de Secuencia Host A Host B TCP Data TCP HDR TCP HDR ISN (initial sequence number) Seq_Num = 1 er byte ACK Seq_Num = Próximo byte esperado

13 13 Initial Sequence Numbers Connection Setup 3-way handshake (Active) Client (Passive) Server Syn +ISN A Syn + Ack +ISN B Ack

14 14 TCP Sliding Window  How much data can a TCP sender have outstanding in the network?  How much data should TCP retransmit when an error occurs? Just selectively repeat the missing data?  How does the TCP sender avoid over-running the receiver’s buffers?

15 15 TCP Sliding Window Window Size Outstanding Un-ack’d data Data OK to send Data not OK to send yet Data ACK’d  Retransmission policy is “Go Back N”.  Current window size is “advertised” by receiver (usually 4k – 8k Bytes when connection set-up).

16 16 TCP Sliding Window Host A Host B ACK Window Size Round-trip time (1) RTT > Window size ACK Window Size Round-trip time (2) RTT = Window size ACK Window Size ???

17 17 Jacobson ( 1988)

18 18 TCP: Retransmission and Timeouts Host A Host B ACK Round-trip time (RTT) ACK Retransmission TimeOut (RTO) Estimated RTT Data1Data2 Guard Band TCP uses an adaptive retransmission timeout value: Congestion Changes in Routing RTT changes frequently

19 19 TCP: Retransmission and Timeouts Picking the RTO is important:  Pick a values that’s too big and it will wait too long to retransmit a packet,  Pick a value too small, and it will unnecessarily retransmit packets. The original algorithm for picking RTO: 1. EstimatedRTT =  EstimatedRTT + (1 -  ) SampleRTT 2. RTO = 2 * EstimatedRTT Characteristics of the original algorithm:  Variance is assumed to be fixed.  But in practice, variance increases as congestion increases.

20 20 TCP Timer Management Of the several timers TCP maintains the most important is the retransmission timer RTO, (also called timeout). After each segment is sent, TCP starts a retransmission timer, if ACK arrives before timer expires, cancel timer. If timer expires first, consider segment lost. How long should RTO be ? Typically some small multiple of RTT. So how to measure RTT ? Measure time between segment sent and ACK receiver. Unfortunately, in the Internet RTT are not constant, they a vary a lot.

21 21 TCP Timer Management (a) Probability density of ACK arrival times in the data link layer. (b) Probability density of ACK arrival times for TCP.

22 22 Retransmisión Adaptiva (Algoritmo Original) Mide SampleRTT para cada par segmento/ ACK Calcula el promedio ponderado de RTT –EstimatedRTT =  x EstimatedRTT +  x SampleRTT –donde  +  = 1  <= 0.9  <= 0.2 Fijar timeout basado en EstimatedRTT –TimeOut = 2 x EstimatedRTT

23 23 Algoritmo de Karn/Partridge No considerar RTT cuando se retransmite Duplicar timeout luego de cada retransmisión SenderReceiver Original transmission ACK SampleR TT Retransmission SenderReceiver Original transmission ACK SampleR TT Retransmission

24 24 Algoritmo de Jacobson/ Karels Nueva forma de calcular el promedio de RTT Diff = sampleRTT - EstRTT EstRTT = EstRTT + (  x Diff) Dev = Dev +  ( |Diff| - Dev) –donde  es un factor entre 0 y 1 (Por ejemplo 1/8) Considerar varianza cuando fijamos el timeout TimeOut =  x EstRTT +  x Dev –donde  = 1 y  = 4 Notas –Los algoritmos son tan buenos/malos como la granularidad del reloj (500ms en Unix) –Un preciso mecanismo de timeout es importante para controlar la congestión (más adelante) –Además de controlar congestión, la idea es no retransmitir cuando no es necesario.

25 25 RTO ( Ret. Timeout) exceptions Assume a segment times out and is then retransmitted. An ACK for the segment arrives. So for purposes for calculating M how do we decide if the ack is for the first send or the retransmission ? We cannot. It might be for the first, but very delayed, or might be for the second. So we cannot use ACKs of retransmitted segments for calculating M (or updating RTT). Rule: Don't use acks of retransmitted segments to update RTT. Instead, if segment times out, simply double RTO. This is called the Karn's algorithm.

26 26 Ejemplo de estimación de RTT:

27 27 TCP 2 parte

28 28 TCP Repaso ( basado en el Peterson)

29 29 Protocolos End-to-End Se apoyan en la capa Red, la cual es de mejor esfuerzo (best-effort) –descarta mensajes –re-ordena mensajes –puede entregar múltiples copias de un mensaje dado –limita los mensajes a algún tamaño finito –entrega mensajes después de un tiempo arbitrariamente largo Servicios comunes ofrecidos/deseados end-to-end –garantía de entrega de mensajes –entrega de mensajes en el mismo orden que son enviados –entrega de a lo más una copia de cada mensaje –soporte para mensajes arbitrariamente largos mensajes –soporte de sincronización –permitir al receptor controlar el flujo de datos del transmisor –soportar múltiples procesos de nivel aplicación en cada máquina

30 30 Demultiplexor Simple (UDP: User Datagram Protocol) Servicio de entrega no confiable y no ordenado de datagramas Agrega multiplexión No hay control de flujo Hay puertos definidos en cada extremo –servidor posee un puerto bien conocido –ver /etc/services en Unix (o Linux) Formato de encabezado Chequeo se suma opcional –psuedo header(IP) + UDP header + data SrcPortDstPort ChecksumLength Data 01631

31 VersionHLen TOSLength IdentFlagsOffset TTLProtocolChecksum SourceAddr DestinationAddr Options (variable) Pad (variable) 048161931 Contexto para encabezado UDP SrcPortDstPort ChecksumLength Data IP UDP Pseudo encabezado Largo de encabezado + datos

32 32 TCP Generalidades Orientado a conexión flujo de byte –app escriben bytes –TCP envía segmentos –app lee bytes Application process Write bytes TCP Send buffer Segment Transmit segments Application process Read bytes TCP Receive buffer … …… Full duplex Control de flujo: evita que el Tx rebalse al receptor Control de congestión: evita que el Tx sobrecargue la red

33 33 Enlace de Datos Versus Transporte Potencialmente conecta muchas máquinas diferentes –requiere de establecimiento y término de conexión explícitos Potencialmente diferentes RTT –requiere mecanismos adaptivos para timeout Potencialmente largos retardos en la red –requiere estar preparado par el arribo de paquetes muy antiguos Potencialmente diferente capacidad en destino –requiere acomodar diferentes capacidades de nodos Potencialmente diferente capacidad de red –requiere estar preparado para congestión en la red

34 34 VersionHLen TOSLength IdentFlagsOffset TTLProtocolChecksum SourceAddr DestinationAddr Options (variable) Pad (variable) 048161931 Contexto Formato de Segmento Options (variable) Data Checksum SrcPortDstPort HdrLen 0Flags UrgPtr AdvertisedWindow SequenceNum Acknowledgment IP TCP Pseudo encabezado URG|ACK|PSH|RST|SYN|FIN

35 35 Formato de Segmento

36 36 Formato de Segmento (cont) Cada conexión es identificada por la 4-tupla: –(SrcPort, SrcIPAddr, DsrPort, DstIPAddr) Ventana deslizante + control de flujo –acknowledgment, SequenceNum, AdvertisedWinow Flags –SYN, FIN, RESET, PUSH, URG, ACK Checksum –pseudo header(IP) + TCP header + data Sender Data(SequenceNum) Acknowledgment + AdvertisedWindow Receiver

37 37 Establecimiento y Término de Conexión Active participant (client) Passive participant (server) SYN, SequenceNum = x SYN + ACK, SequenceNum = y, ACK, Acknowledgment = y + 1 Acknowledgment = x + 1

38 38 Diagrama de Estado de Transmisión CLOSED LISTEN SYN_RCVDSYN_SENT ESTABLISHED CLOSE_WAIT LAST_ACKCLOSING TIME_WAIT FIN_WAIT_2 FIN_WAIT_1 Passive openClose Send/SYN SYN/SYN + ACK SYN + ACK/ACK SYN/SYN + ACK ACK Close/FIN FIN/ACKClose/FIN FIN/ACK ACK + FIN/ACK Timeout after two segment lifetimes FIN/ACK ACK Close/FIN Close CLOSED Active open/SYN

39 39 Revisión de Ventana Deslizante Lado Transmisor LastByteAcked < = LastByteSent LastByteSent < = LastByteWritten Se tiene en buffer los bytes entre LastByteAcked y LastByteWritten Sending application LastByteWritten TCP LastByteSentLastByteAcked Receiving application LastByteRead TCP LastByteRcvdNextByteExpected Lado Receptor LastByteRead < NextByteExpected NextByteExpected < = LastByteRcvd +1 Se tiene en buffer los bytes entre NextByteRead y LastByteRcvd LastByteRead+1

40 40 Control de Flujo Tamaño del buffer de envío: MaxSendBuffer Tamaño del buffer de recepción: MaxRcvBuffer Lado receptor –LastByteRcvd - LastByteRead < = MaxRcvBuffer –AdvertisedWindow = MaxRcvBuffer - (LastByteRcvd - NextByteRead) Lado Transmisor –LastByteSent - LastByteAcked < = AdvertisedWindow –EffectiveWindow = AdvertisedWindow - (LastByteSent - LastByteAcked) –LastByteWritten - LastByteAcked < = MaxSendBuffer –Bloquear Tx si (LastByteWritten - LastByteAcked) + y > MaxSenderBuffer, y bytes que se desean escribir. Siempre enviar ACK en respuesta a la llegada de segmentos de datos Tx persiste enviando 1 byte cuando AdvertisedWindow = 0 Sending application LastByteWritten TCP LastByteSentLastByteAcked Receiving application LastByteRead TCP LastByteRcvdNextByteExpected

41 41 ¿Qué tan agresivamente el Tx explota la apertura de ventana? Soluciones en lado Receptor –Retardar los acuses de recibo Síndrome de Ventana estúpida (Silly) SenderReceiver

42 42 Silly Window Syndrome

43 43 Algoritmo de Nagle ¿Qué tanto tiempo el Tx retarda la transmisión de datos? –Demasiado largo: afecta aplicaciones interactivas –Demasiado corto: Utilización de la red es pobre –Estrategias: Basadas en timers v/s auto relojes Cuando la aplicación genera datos adicionales: –Si se llena un segmento (y la ventana está abierta): enviar –Sino Si hay datos sin ack en Tx: dejar en buffer hasta llegada de ack sino: enviar datos

44 44 Nagle's algorithm Purpose is to allow the sender TCP to make efficient use of the network, while still being responsive to the sender applications. Idea: If application data comes in byte by byte, send first byte only. Then buffer all application data till until ACK for first byte comes in. If network is slow and application is fast, the second segment will contain a lot of data. Send second segment and buffer all data till ACK for second segment comes in. This way the algorithm is clocking the sends to speed of the network and simultaneously preventing sending several one byte segments back to back. An exception to this rule is to always send (not wait for ACK) if enough data for half the receiver window or MSS.

45 45 Protección contra reapariciones de igual número de secuencia SequenceNum de 32 bits BandwidthTiempo hasta tener problema T1 (1.5 Mbps)6.4 hours Ethernet (10 Mbps)57 minutes T3 (45 Mbps)13 minutes FDDI (100 Mbps)6 minutes STS-3 (155 Mbps)4 minutes STS-12 (622 Mbps)55 seconds STS-24 (1.2 Gbps)28 seconds

46 46 Mantención de la tubería llena AdvertisedWindow de 16 bits BandwidthDelay x Bandwidth Product T1 (1.5 Mbps)18KB Ethernet (10 Mbps)122KB T3 (45 Mbps)549KB FDDI (100 Mbps)1.2MB STS-3 (155 Mbps)1.8MB STS-12 (622 Mbps)7.4MB STS-24 (1.2 Gbps)14.8MB Asumiendo RTT de 100 ms 64 KB

47 47 Extensiones de TCP Son implementadas como opciones del encabezado Almacenar marcas de tiempo en segmentos de salida Extender espacio de secuencia con marca de tiempo de 32-bit (PAWS) Desplazar (escalar) ventana avisada. La idea es medir la ventana en unidades de 2, 4, 8 bytes.

48 48 TCP Options Some TCP options are: Maximum segment size (MSS): Specified what is the payload the sender is able to receive. (Default MSS = 536 bytes, i.e., Segment size = MSS + 20). SMSS/RMSS is Sender/Receiver MSS. Window scale: The window size field allows for upto 2^16 bytes of data. But this might be inefficient for high bw x delay situations. This options TCP indicate a scaling factor. Negative acknowledgement: Lets receiver user NAKs to get realize selective repeat rather than the normal go-back-N TCP behaviour.

49 49 Recordemos MSS Application Message TCP dataTCP hdr MSS TCP Segment IP dataIP hdr IP Packet Ethernet dataEthernet Ethernet Frame 20 bytes 14 bytes 4 bytes MTU 1500 bytes

50 50 Checksum Se calcula entre el TCP Header, Data y el pseudo header

51 51 Window Management

52 52 Initial Sequence Number Select initial sequence numbers (ISN) to protect against segments from prior connections (that may circulate in the network and arrive at a much later time) Select ISN to avoid overlap with sequence numbers of prior connections Use local clock to select ISN sequence number Time for clock to go through a full cycle should be greater than the maximum lifetime of a segment (MSL); Typically MSL=120 seconds

53 53 TCP Connection Establishment ( a) TCP connection establishment in the normal case. (b) Call collision. 6-31 Initial sequence numbers are not 0. TCP uses a clock tick counter (at 4 usecs rate) to setup the initial sequence numbers. This scheme prevents delayed duplicates.

54 54 Connection Establishment (cont) Active participant (client) Passive participant (server) SYN, SequenceNum = x ACK, Acknowledgment =y+1 Acknowledgment =x+1 SYN+ACK, SequenceNum=y,

55 55 TCP Connection Release Graceful release: –Each side of the connection released independently. Either side send TCP segment with FIN=1. When FIN acknowledged, that direction is shut down for data. Connection released when both sides shut down. –4 segments: 1 FIN and 1 ACK for each direction; 1st. ACK+2nd. FIN combined. –Two-army problem, Timers, 2 MSL

56 56 TCP Connection Management Modeling TCP connection management finite state machine. The heavy solid line is the normal path for a client. The heavy dashed line is the normal path for a server. The light lines are unusual events. Each transition is labeled by the event causing it and the action resulting from it, separated by a slash.

57 57 TCP Connection Management Modeling The states used in the TCP connection management finite state machine.

58 58 Maximum Segment Size –largest block of data that TCP sends to other end Each end can announce its MSS during connection establishment Default is 576 bytes including 20 bytes for IP header and 20 bytes for TCP header Ethernet implies MSS of 1460 bytes IEEE 802.3 implies 1452

59 59 Tear-down Packet Exchange SenderReceiver FIN FIN-ACK FIN FIN-ACK Data write Data ack

60 60 Connection Tear-down

61 61 Detecting Half-open Connections

62 62 TIME-WAIT Assassination