Download 3rd Edition: Chapter 3

1DT057
Distributed Information System
Transport Layer
1
2: Application Layer
Chapter 3
CHAPTER 3: TRANSPORT LAYER




multiplexing/demultipl
exing
reliable data transfer
flow control
congestion control

learn about transport
layer protocols in the
Internet:


UDP: connectionless
transport
TCP: connection-oriented
transport
Transport Layer
Our goals:
 understand principles
behind transport
layer services:
3-2
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
Congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

3-3
TRANSPORT VS. NETWORK LAYER
network layer: logical
communication
between hosts
 transport layer: logical
communication
between processes

Transport Layer

relies on, enhances,
network layer services
3-4
INTERNET TRANSPORT-LAYER PROTOCOLS

reliable, in-order
delivery (TCP)



unreliable, unordered
delivery: UDP


congestion control
flow control
connection setup
no-frills extension of “besteffort” IP
services not available:


delay guarantees
bandwidth guarantees
network
data link
physical
network
data link
physical
network
data link
physicalnetwork
network
data link
physical
Transport Layer

application
transport
network
data link
physical
data link
physical
network
data link
physical
application
transport
network
data link
physical
3-5
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
Congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

3-6
MULTIPLEXING/DEMULTIPLEXING
Multiplexing at send host:
Demultiplexing at rcv host:
gathering data from multiple
sockets, enveloping data with
header (later used for
demultiplexing)
delivering received segments
to correct socket
= socket
application
transport
network
link
= process
P3
P1
P1
application
P2
P4
application
transport
transport
network
network
link
link
physical
host 1
physical
physical
host 3
host 2
Transport Layer
3-7
Hhost
OW
DEMULTIPLEXING WORKS
receives IP datagrams


32 bits
source port #
dest port #
other header fields
Transport Layer
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

application
data
(message)
TCP/UDP segment format
3-8
CONNECTIONLESS DEMUX (CONT)
DatagramSocket serverSocket = new DatagramSocket(6428);
SP: 6428
SP: 6428
DP: 9157
DP: 5775
SP: 9157
client
IP: A
P1
P1
P3
DP: 6428
SP provides “return address”
Transport Layer
P2
SP: 5775
server
IP: C
DP: 6428
Client
IP:B
3-9
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

310
UDP: USER DATAGRAM PROTOCOL [RFC
768]


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

“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
311
UDP: MORE

other UDP uses
DNS
 SNMP
reliable transfer over UDP:
add reliability at application
layer
 application-specific error
recovery!


segment,
including
header
32 bits
source port #
dest port #
length
checksum
Transport Layer

often used for streaming
multimedia apps
 loss tolerant
Length, in
bytes of UDP
 rate sensitive
Application
data
(message)
UDP segment format
312
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

313
PRINCIPLES OF RELIABLE DATA
TRANSFER


important in app., transport, link layers
top-10 list of important networking topics!
Transport Layer

characteristics of unreliable channel will determine complexity of reliable data transfer
protocol (rdt)
314
PRINCIPLES OF RELIABLE DATA
TRANSFER


important in app., transport, link layers
top-10 list of important networking topics!
Transport Layer

characteristics of unreliable channel will determine complexity of reliable data transfer
protocol (rdt)
315
PRINCIPLES OF RELIABLE DATA
TRANSFER


important in app., transport, link layers
top-10 list of important networking topics!
Transport Layer

characteristics of unreliable channel will determine complexity of reliable data transfer
protocol (rdt)
316
RDT3.0 EXAMPLE
Transport Layer
317
RDT3.0 EXAMPLE
Transport Layer
318
PERFORMANCE OF RDT3.0
rdt3.0 works, but performance stinks
 ex: 1 Gbps link, 15 ms prop. delay, 8000 bit packet:

Transport Layer
L 8000bits
d trans  
 8 microsecon ds
9
R 10 bps

U sender: utilization – fraction of time sender busy sending
U


sender
=
L/R
RTT + L / R
=
.008
30.008
= 0.00027
microsec
onds
1KB pkt every 30 msec -> 33kB/sec thruput over 1 Gbps link
network protocol limits use of physical resources!
319
RDT3.0: STOP-AND-WAIT OPERATION
sender
receiver
first packet bit transmitted, t = 0
last packet bit transmitted, t = L / R
Transport Layer
first packet bit arrives
last packet bit arrives, send ACK
RTT
ACK arrives, send next
packet, t = RTT + L / R
U
=
sender
L/R
RTT + L / R
=
.008
30.008
= 0.00027
microsec
onds
320
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
Transport Layer

Two generic forms of pipelined protocols: go-Back-N,
3selective repeat
21
PIPELINING: INCREASED UTILIZATION
sender
receiver
first packet bit transmitted, t = 0
last bit transmitted, t = L / R
RTT
Transport Layer
first packet bit arrives
last packet bit arrives, send ACK
last bit of 2nd packet arrives, send ACK
last bit of 3rd packet arrives, send ACK
ACK arrives, send next
packet, t = RTT + L / R
Increase utilization
by a factor of 3!
U
sender
=
3*L/R
RTT + L / R
=
.024
30.008
= 0.0008
microsecon
ds
3-22
PIPELINING PROTOCOLS


doesn’t ACK pkt if
there’s a gap
sender: has timer for
oldest unACKed pkt

if timer expires:
retransmit all
unACKed packets
Selective Repeat: overview
 sender: up to N unACKed
packets in pipeline
 receiver: ACKs individual
pkts
 sender: maintains timer for
each unACKed pkt
Transport Layer
Go-back-N: overview
 sender: up to N
unACKed pkts in
pipeline
 receiver: only sends
cumulative ACKs

if timer expires: retransmit
only unACKed packet
323
GO-BACK-N
Sender:

Transport Layer

k-bit seq # in pkt header
“window” of up to N, consecutive unACKed 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
3 timeout(n): retransmit pkt n and all higher seq # pkts in window

24
GBN IN
ACTION
Transport Layer
325
GBN DEMO
Transport Layer
http://www.cs.mum.edu/courses/cs450/GoBackN.htm
326
SELECTIVE REPEAT

receiver individually acknowledges all correctly
received pkts

sender only resends pkts for which ACK not
received


buffers pkts, as needed, for eventual in-order delivery to
upper layer
Transport Layer

sender timer for each unACKed pkt
sender window


N consecutive seq #’s
again limits seq #s of sent, unACKed pkts
327
SELECTIVE REPEAT: SENDER, RECEIVER
WINDOWS
Transport Layer
328
SELECTIVE REPEAT
sender
data from above :

timeout(n):

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 #
 send ACK(n)
 out-of-order: buffer
Transport Layer
if next available seq # in
window, send pkt
receiver
pkt n in [rcvbase, rcvbase+N-1]
 in-order: deliver (also
deliver buffered, in-order
pkts), advance window to
next not-yet-received pkt
pkt n in
[rcvbase-N,rcvbase-1]
 ACK(n)
otherwise:
 ignore
329
SELECTIVE REPEAT IN ACTION
Transport Layer
330
SELECTIVE REPEAT:
DILEMMA
Example:



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

seq #’s: 0, 1, 2, 3
window size=3
331
SELECTIVE REPEAT DEMO
Transport Layer
http://www.eecis.udel.edu/~amer/450/TransportApplets/
SR/SRindex.html
332
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
Congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

333
TCP: OVERVIEW

socket
door
point-to-point:
reliable, in-order byte
steam:

pipelined:

send & receive buffers
application
writes data
application
reads data
TCP
send buffer
TCP
receive buffer
segment

full duplex data:

connection-oriented:

flow controlled:
socket
door
Transport Layer

RFCS: 793, 1122, 1323, 2018, 2581
334
TCP SEQ. #’S AND ACKS
Host A
User
types
‘C’
Host B
Transport Layer
Seq. #’s:
 byte stream
“number” of first
byte in segment’s
data
ACKs:
 seq # of next byte
expected from other
side
 cumulative ACK
host ACKs
receipt of
‘C’, echoes
back ‘C’
host ACKs
receipt
of echoed
‘C’
simple telnet scenario
time
335
TCP RELIABLE DATA TRANSFER

retransmissions are
triggered by:


timeout events
duplicate ACKs
Transport Layer
TCP creates rdt
service on top of IP’s
unreliable service
 pipelined segments
 cumulative ACKs
 TCP uses single
retransmission timer

336
TCP: RETRANSMISSION SCENARIOS
Host A
X
loss
Sendbase
= 100
SendBase
= 120
SendBase
= 100
time
SendBase
= 120
lost ACK scenario
Host B
Seq=92 timeout
Seq=92 timeout
Host B
Transport Layer
timeout
Host A
time
3premature timeout
37
TCP RETRANSMISSION SCENARIOS (MORE)
Host B
X
loss
Transport Layer
timeout
Host A
SendBase
= 120
time
Cumulative ACK scenario
338
TCP FLOW CONTROL

flow control
receive side of TCP
connection has a
receive buffer:
(currently)
TCP data application
IP
unused buffer
(in buffer)
process
datagrams
space

Transport Layer
sender won’t overflow
receiver’s buffer by
transmitting too much,
too fast
speed-matching service:
matching send rate to
receiving application’s
drain rate
 app process may be
slow at reading from
buffer
339
CHAPTER 3 OUTLINE

3.5 Connection-oriented
transport: TCP




segment structure
reliable data transfer
flow control
congestion control
Transport Layer
3.1 Transport-layer
services
 3.2 Multiplexing and
demultiplexing
 3.3 Connectionless
transport: UDP
 3.4 Principles of
reliable data transfer

340
PRINCIPLES OF CONGESTION CONTROL
Transport Layer
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)
 long delays (queueing in router buffers)

341
TCP CONGESTION CONTROL:
 goal: TCP sender should transmit as fast as possible,
but without congesting network
Q: how to find rate just below congestion level
 decentralized: each TCP sender sets its own rate,
Transport Layer

based on implicit feedback:
 ACK: segment received (a good thing!), network not
congested, so increase sending rate
 lost segment: assume loss due to congested
network, so decrease sending rate
342
TCP CONGESTION CONTROL: BANDWIDTH
PROBING
 “probing for bandwidth”: increase transmission rate
continue to increase on ACK, decrease on loss (since available
bandwidth is changing, depending on other connections in
network)
ACKs being received,
so increase rate
sending rate

Transport Layer
on receipt of ACK, until eventually loss occurs, then
decrease transmission rate
X
X loss, so decrease rate
X
X
TCP’s
“sawtooth”
behavior
X
time
343
CHAPTER 3: SUMMARY

principles behind transport
layer services:

instantiation and
implementation in the
Internet
UDP
 TCP

Next:
 leaving the network
“edge” (application,
transport layers)
 into the network
“core”
Transport Layer
multiplexing, demultiplexing
 reliable data transfer
 flow control
 congestion control

344