1. Data Communication and Networks
Lecture 11
Network Congestion:
Causes, Effects, Controls
November 16, 2006
Transport Layer

2. What Is Congestion?
Congestion occurs when the number of packets being transmitted through the network approaches the packet handling capacity of the network
Congestion control aims to keep number of packets below level at which performance falls off dramatically
Data network is a network of queues
Generally 80% utilization is critical
Finite queues mean data may be lost
A top-10 problem!
Transport Layer

3. Queues at a Node
Transport Layer

4. Effects of Congestion Packets arriving are stored at input buffers
Routing decision made
Packet moves to output buffer
Packets queued for output transmitted as fast as possible
Statistical time division multiplexing
If packets arrive to fast to be routed, or to be output, buffers will fill
Can discard packets
Can use flow control
Can propagate congestion through network
Transport Layer

5. Interaction of Queues
Transport Layer

6. Causes/costs of congestion: scenario 1
two senders, two receivers
one router, infinite buffers
no retransmission
large delays when congested
maximum achievable throughput
Hosts A and B both send at rate in
Packets pass through router and over shared link of capacity C
Router has buffers to store outgoing packets
Clearly, per connection rate of delivery from router to destination (from A or B) cannot exceed C/2, no matter what the value of in is. That is, out <= C/2
Note that when in > C/2, delay grows asymptotically since queue depth in router is infinite.
Transport Layer

7. Causes/costs of congestion: scenario 2
one router, finite buffers
sender retransmission of lost packet
Router in this case has a finite number of buffers.
Assume transport layer is “reliable”: that is, if router drops a packet, sender (A or B) will retransmit.
“Offered load” now includes both “original” packets (rate is in) and “retransmitted” packets.
The rate for all packets transmitted by A or B (original + retransmitted) is ’in.
Transport Layer

8. Causes/costs of congestion: scenario 2l in
out =
always: (’in = in)
“perfect” retransmission only when loss:
retransmission of delayed (not lost) packet makes larger (than perfect case) for same l in
out
> l in l
out
Performance depends on how A and B do retransmission.
Suppose A,B send only when they know router has a free buffer, so no loss occurs. Then, ’in = in.
Somewhat more realistically, suppose A,B only retransmit when know for sure that packet is lost (“perfect retransmission”). Then packets arrive at destination at a rate < C/2 (middle graph is an example).
Even more realistically, some packets will be sent twice because sender’s timer is occasionally not quite long enough. So, delivery rate to receiver is even worse since receiver will discard duplicate (right hand graph).
So, cost of congestion is BOTH cost of resend lost packet AND cost of duplicates (wasted transmission).
“costs” of congestion:
more work (retrans) for given “goodput”
unneeded retransmissions: link carries multiple copies of pkt
Transport Layer

9. Causes/costs of congestion: scenario 3
four senders
multihop paths
timeout/retransmit l in
Q: what happens as and increase ? l in
A-C path shares router R1 with D-B path, and shares R2 with B-D path
Connections between routers have transmission capacity C.
When in is small, no retransmission occurs.
Transport Layer

10. Causes/costs of congestion: scenario 3
Consider this case:
Since R2 is second hop for A-C, packets on A-C path from R1 cannot arrive at R2 at a rate > C, regardless of rate of in.
If ’in is very large for all connections, then arrival rate of B packets at R2 can be >> than A packets sent from R1.
So, more of R2’s buffers are used for B packets than A packets. As offered load goes even higher, eventually only B packets are in buffer There is no room for A packets, so R2 drops A packets.
So, throughput on A-C path (measured at C) goes to ZERO!
Another cost of congestion: When a packet is dropped at a downstream node on a path, all of the capacity used to send the packet from the upstream nodes has been WASTED!
Another “cost” of congestion:
when packet dropped, any “upstream transmission capacity used for that packet was wasted!
Transport Layer

11. Approaches towards congestion control
Two broad approaches towards congestion control:
End-end congestion control:
no explicit feedback from network
congestion inferred from end-system observed loss, delay
approach taken by TCP
Network-assisted congestion control:
routers provide feedback to end systems
single bit indicating congestion (SNA, DECbit, TCP/IP ECN, ATM)
explicit rate sender should send at
Transport Layer

12. Case study: ATM ABR congestion control
ABR: available bit rate:
“elastic service”
if sender’s path “underloaded”:
sender should use available bandwidth
if sender’s path congested:
sender throttled to minimum guaranteed rate
RM (resource management) cells:
sent by sender, interspersed with data cells
bits in RM cell set by switches (“network-assisted”)
NI bit: no increase in rate (mild congestion)
CI bit: congestion indication
RM cells returned to sender by receiver, with bits intact
Transport Layer

13. Case study: ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell
congested switch may lower ER value in cell
sender’ send rate thus minimum supportable rate on path
EFCI bit in data cells: set to 1 in congested switch
if data cell preceding RM cell has EFCI set, sender sets CI bit in returned RM cell
Transport Layer

14. TCP Congestion Control
end-end control (no network assistance)
sender limits transmission:
LastByteSent-LastByteAcked
 CongWin
Roughly,
CongWin is dynamic, function of perceived network congestion
How does sender perceive congestion?
loss event = timeout or 3 duplicate acks
TCP sender reduces rate (CongWin) after loss event
three mechanisms:
AIMD
slow start
conservative after timeout events
rate =
CongWin
RTT
Bytes/sec
Transport Layer

15. TCP AIMD multiplicative decrease: cut CongWin in half after loss event
additive increase: increase CongWin by 1 MSS every RTT in the absence of loss events: probing
Long-lived TCP connection
Transport Layer

16. TCP Slow Start
When connection begins, increase rate exponentially fast until first loss event
When connection begins, CongWin = 1 MSS
Example: MSS = 500 bytes & RTT = 200 msec
initial rate = 20 kbps
available bandwidth may be >> MSS/RTT
desirable to quickly ramp up to respectable rate
Transport Layer

17. TCP Slow Start (more)
When connection begins, increase rate exponentially until first loss event:
double CongWin every RTT
done by incrementing CongWin for every ACK received
Summary: initial rate is slow but ramps up exponentially fast
Host A
Host B
one segment
RTT
two segments
four segments
time
Transport Layer

18. Refinement
Philosophy:
3 dup ACKs indicates network capable of delivering some segments
timeout before 3 dup ACKs is “more alarming”
After 3 dup ACKs:
CongWin is cut in half
window then grows linearly
But after timeout event:
CongWin instead set to 1 MSS;
window then grows exponentially
to a threshold, then grows linearly
Transport Layer

19. Refinement (more) Implementation:
Q: When should the exponential increase switch to linear?
A: When CongWin gets to 1/2 of its value before timeout.
Implementation:
Variable Threshold
At loss event, Threshold is set to 1/2 of CongWin just before loss event
Transport Layer

20. Summary: TCP Congestion Control
When CongWin is below Threshold, sender in slow-start phase, window grows exponentially.
When CongWin is above Threshold, sender is in congestion-avoidance phase, window grows linearly.
When a triple duplicate ACK occurs, Threshold set to CongWin/2 and CongWin set to Threshold.
When timeout occurs, Threshold set to CongWin/2 and CongWin is set to 1 MSS.
Transport Layer

21. TCP Fairness
Fairness goal: if K TCP sessions share same bottleneck link of bandwidth R, each should have average rate of R/K
TCP connection 1
bottleneck
router
capacity R
TCP
connection 2
Transport Layer

22. Why is TCP fair? Two competing sessions:
Additive increase gives slope of 1, as throughout increases
multiplicative decrease decreases throughput proportionally R
equal bandwidth share
loss: decrease window by factor of 2
congestion avoidance: additive increase
Connection 2 throughput
loss: decrease window by factor of 2
congestion avoidance: additive increase
Connection 1 throughput R
Transport Layer

23. Fairness (more) Fairness and parallel TCP connections Fairness and UDP
nothing prevents app from opening parallel cnctions between 2 hosts.
Web browsers do this
Example: link of rate R supporting 9 cnctions;
new app asks for 1 TCP, gets rate R/10
new app asks for 11 TCPs, gets R/2 !
Fairness and UDP
Multimedia apps often do not use TCP
do not want rate throttled by congestion control
Instead use UDP:
pump audio/video at constant rate, tolerate packet loss
Research area: TCP friendly
Transport Layer