Principles of
Congestion
Control
Chapter 3.6
Computer Networking: A top-down
approach
Principles of Congestion Control
Congestion:
• informally: “too many sources sending too much
data too fast for network to handle”
• manifestations:
o lost packets (buffer overflow at routers)
o long delays (queuing in router buffers)
• different from flow control!
• a top-10 problem!
• 3 examples of cause and costs of congestion
control
Transport Layer
3-2
Causes/costs of congestion: scenario 1
• two senders, two
receivers
• one router,
infinite buffers
• no retransmission
Host A
Host B
lout
lin : original data
Host C
unlimited shared
output link buffers
Host D
• large delays
when
congested
• maximum
achievable
throughput
Transport Layer
3-3
Causes/costs of congestion: scenario 2
• one router, finite buffers
• sender retransmission of lost packet
Host A
Host B
Transport Layer
lin : original
data
l'in : original data, plus
retransmitted data
lout
Host C
finite shared output
link buffers
3-4
Causes/costs of congestion: scenario 2
= l
(goodput)
out
in
• “perfect” retransmission only when loss:
• always:
l
l > lout
in
• retransmission of delayed (not lost) packet makes
(than perfect case) for same
C/2
lout
C/2
l
in
larger
C/2
lin
a.
C/2
lout
lout
lout
C/3
lin
b.
C/2
C/4
lin
C/2
c.
“costs” of congestion:
 more work needed for given “goodput” (retransmission)
 unneeded retransmissions: link carries multiple copies of packets
Transport Layer
3-5
Causes/costs of congestion: scenario 3
• four senders
• multihop paths
• timeout/retransmit
lin : original data
Host A
lout
l'in : original data, plus
retransmitted data
R1
Host B
finite shared
output link
buffers
Host D
R2
Transport Layer
Host C
3-6
Causes/costs of congestion: scenario 3
Host A
lout
Host B
Another “cost” of congestion:
 when packet dropped, any “upstream transmission
capacity used for that packet was wasted!
Transport Layer
3-7
Approaches towards congestion control
Two broad approaches towards congestion control:
End-to-end congestion
control:
Network-assisted
congestion control:
• no explicit feedback from
network
• congestion inferred from
end-system observed loss,
delay
• approach taken by TCP
• routers provide feedback
to end systems
o single bit indicating
congestion (SNA,
DECbit, TCP/IP ECN,
ATM)
o explicit rate sender
should send at
Transport Layer
3-8
Network-assisted congestion control
• Feedback in two ways
o Direct feedback – network router  sender (choke packet)
o Network feedback via receiver – Router marks a packet and receiver notify
sender