ch3 e1 3600

Problem with stop and wait protocols

ACKing every packet will take a long time

NACK-less channel approach

sender waits “reasonable” amount of time for an ACK. It will retransmit if the ACK is not received in time

If packet or ACK is delayed and not lost retransmission will be a

duplicate

If retransmission pkt is a duplicate, the ____ will handle this

sequence number

receiver must specify ______ _______ of pkt being ACKed

sequence number

pipelining

sender allows multiple, “in-flight”, yet-to-be-acknowledged pkts with seq#

two forms of pipelined protocols

go-back-N (GBN), selective repeat

3-packet pipelining formula

(3L / R) / (RTT + (L/R))

Go-Back-N

sender can have up to N unacked packets in pipeline

GBN: receiver only send ______ _____

cumulative ack

Selective Repeat: receiver sends _____ ____ for each packet

individual ack

Go-Back-N: sender

k-bit sequence number in packet header

ACK(n):

ACKs all pkts up to, including seq # n -“cumulative ACK”

timeout(n)

retransmit packet n and all higher seq # pkts in window

GBN receiver Ack-only:

always send ACK for correctly received pkt with highest in-order seq #

out-of-order pkt

discard, dont buffer, re-ACK packet with highest in order seq #

Selective repeat

receiver individually acknowledges all correctly received packets

selective repeat: sender only resends pkts for ______

which ACK wasn’t received

Selective repeat: sender window

N consecutive seq #s, limits seq #s of sent unACKED pkts

TCP: point-to-point

one sender, one receiver. no multicasting, NOT circuit-switched

TCP provides

reliable in-order bytes-stream

TCP: pipelined

TCP congestion and flow set the window size

TCP: full duplex data

bi-directional data flow in same connection

MSS

maximum segment size

TCP: connection-oriented

handshaking initializes the sender-receiver state before data exchange

TCP: flow controlled

sender will not overwhelm receiver

TCP Sequence Number

Indicate the position in the byte stream, specifically the number of the first byte of data carried in the segment.

TCP Acknowledgements

Sequence # of next byte expected from other side. cumulative ack

how does the receiver handle out-of-order segments?

TCP spec doesn’t say, up to implementor

How to estimate RTT?

SampleRTT: measured time from segment transmission until ACK receipt

To get a more accurate RTT, take _______ measurements

multiple, then average

TCP reliable data transfer

TCP creates rdt service on top of IPs unreliable service

retransmissions are triggered by

timeout events, duplicate acks

TCP sender events: data received from app

Create segment with seq #, seq # is byte-stream number of first data byte in segment, start the timer if not already running

TCP sender events: timeout

retransmit segment that caused timeout, restart timer

TCP sender events: ack rcvd

if ack acknowledges previously unacked segments: update what is known to be ACKed, start timer if there are still unacked segments

Event at receiver: arrival of in-order segment with expected seq #. All data up to expected seq # is already ACKed. What is the TCP receiver action?

delayed ACK, wait up to 500ms for next segment. if no next segment, send ACK

Event at receiver: arrival of in-order segment with expected seq #. One other segment has ACK pending. What is the TCP receiver action?

immediately send single cumulative ACK, ACKing both in-order segments

event at receiver: arrival of out-of-order segment higher-than-expect seq. # . Gap detected. What is the TCP receiver action?

immediately send duplicate ACK, indicating seq. # of next expected byte

Event at receiver: arrival of segment that partially or completely fills gap. What is the TCP receiver action?

immediate send ACK, provided that segment starts at lower end of gap

TCP flow control

receiver controls sender, so sender won’t overflow receiver’s buffer by transmitting too much, too fast

TCP flow control: Step 1

receiver “advertises” free buffer space by including rwnd value in TCP header of receiver-to-sender segments

TCP flow control: step 2

sender limits amount of unacked (“in-flight”) data to receiver’s rwnd value

TCP flow control: final step

guarantees receive buffer will not overflow

2-way handshake will always work in network: T/F

False

TCP: Closing a connection

client, server each close their side of connection with FIN bit = 1

TCP: closing a connection step 2

Respond to received FIN with ACK

congestion

too many sources sending too much data too fast for network to handle

congestion is different from fflow control: T/F

True

lost packets means

buffer overflow at routers

long delays means

queueing in router buffers

sender limits transmission equation

LastByteSent - LastByteAcked <= cwnd

TCP sending rate equation

(cwnd / RTT)

cwnd is ______

dynamic

TCP slow start

when connection begins, increase rate exponentially until first loss event

TCP slow start: double cwnd every ____

RTT

TCP slow start: done by incrementing ______ for every ACK received

cwnd

TCP Slow start summary

initial rate is slow but ramps up exponentially fast

loss is indicated by

timeout

cwnd is set to

1 MSS

loss indicated by _ duplicate ACKs

3

TCP ______ always sets cwnd to 1

tahoe

when should the exponential increase switch to linear?

when cwnd gets to 1/2 of its value before timeout.

cwnd

congestion window

TCP fairness

if K TCP sessions share same bottleneck link of bandwidth R, each should have average rate of R/K

Transport services and protocols provides what

logical communication between app processes running on different hosts

transport protocols run in ___ systems

end

send side:

breaks app messages into segments, passes to network layer

rcv side:

reassembles segments into messages, passes to app layer

network layer

logical communication between hosts

transport

logical communication between processes

transport layer relies on and enhances what?

network layer services

UDP only provides ____ to ____ delivery

process to process

TCP has more services, such as:

flow control, seq #, acks, timers and congestion control

extending host-to-host delivery and process-to-process deliver is called transport layer _______ and _______

multiplexing, demultiplexing

multiplexing at sender

handle data from multiple sockets, add transport header (later used for demultiplexing)

demultiplexing at receiver

use header info to deliver received segments to correct socket

Demultiplexing Step 1:

host receives IP datagrams with source and destination IP address

Demultiplexing Step 2:

each datagram carries one transport-layer segment

Demultiplexing Step 3: 

each segment has source, destination port number

Demultiplexing step 4:

host uses IP addresses & port numbers to direct segment to appropriate socket

TCP socket identified by 4-tuple:

Source IP address, source port number, destination IP address, destination port number

demux

receiver uses all four values to direct segment to appropriate socket

server host may support _______ TCP sockets

multiple

web servers have different ________ for each connecting client

sockets

non-persistent HTTP will have _____ _____ for each request

different socket

UDP stands for

User Datagram Protocol

UDP is a ____ _______ service

best effort

UDP segments may be

lost, delivered out of order

Connectionless UDP

no handshaking between UDP sender, receiver

USP uses

streaming multimedia, DNS, SNMP

reliable transfer over UDP:

add reliability at application layer

UDP Segment header

source port #, destination port #, length & checksum. all 16-bit and create a 32 bit header

UDP checksum goal

detect “errors” (e.g., flipped bits) in transmitted segment

UDP Checksum sender step 1:

treat segment contents, including header fields, as sequence of 16-bit integers

UDP checksum sender step 2

checksum: addition (one’s complement sum) of segment contents

UDP Checksum sender step 3:

sender puts checksum value into UDP checksum field

UDP Checksum receiver step 1:

compute checksum of received segment

UDP checksum receiver step 2:

check if computed checksum equals checksum field value

delivering the data in a transport-layer segment to the correct socket is called

demultiplexing