broadband and high speed commu Chapter10.ppt

gh

1. Chapter 10 Congestion Control in Data Networks and Internets
1
Chapter 10
Congestion Control in Data
Networks and Internets

2. Chapter 10 Congestion Control in Data Networks and Internets
2
Introduction
Congestion occurs when number of
packets transmitted approaches network
capacity
Objective of congestion control:
–keep number of packets below level at which
performance drops off dramatically

3. Chapter 10 Congestion Control in Data Networks and Internets
3
Queuing Theory
Data network is a network of queues
If arrival rate > transmission rate
then queue size grows without bound and packet
delay goes to infinity
Even If arrival rate < transmission rate, queue grows
dramatically as the arrival rate approaches the
transmission rate.
A traditional rule of thumb is that when line efficiency
> 80%, the queue grows at an alarming rate and this is
worse for self-similar traffic for lesser efficiency.
Results in large delays and finally buffer overflow.

4. Chapter 10 Congestion Control in Data Networks and Internets
4
Figure 10.1

5. Chapter 10 Congestion Control in Data Networks and Internets
5
At Saturation Point, 2 Strategies
Discard any incoming packet if no buffer
available
Saturated node exercises flow control over
neighbors
–May cause congestion to propagate throughout
network

6. Chapter 10 Congestion Control in Data Networks and Internets
6
Figure 10.2

7. Chapter 10 Congestion Control in Data Networks and Internets
7
Ideal Performance
I.e., infinite buffers, no overhead for packet
transmission or congestion control
Throughput (no of pkts delivered to dstn)
increases with offered load (no of pkts
transmitted by src) until full capacity
Packet delay increases with offered load
approaching infinity at full capacity
Power = throughput / delay
Higher throughput results in higher delay

8. Chapter 10 Congestion Control in Data Networks and Internets
8
Figure 10.3

9. Chapter 10 Congestion Control in Data Networks and Internets
9
Practical Performance
I.e., finite buffers, non-zero packet processing
overhead
With no congestion control, increased load
eventually causes moderate congestion:
throughput increases at slower rate than load
Further increased load causes packet delays to
increase and eventually throughput to drop to
zero
Some examples of this situation –Collisions,
retransmission (of dropped pkts and delayed
pkts (because ACK timers expire))

10. Chapter 10 Congestion Control in Data Networks and Internets
10
Figure 10.4

11. Chapter 10 Congestion Control in Data Networks and Internets
11
Congestion Control
Backpressure
–Close destination and the effect propagates towards the source where
transmission can be halted.
–Works well in ntks where hop-by-hop flow control exists like in
X.25.
Choke Packet
–A pkts generated at a congested node and transmitted back to a
source node to restrict traffic flow.
–Choke packet: ICMP Source Quench
Implicit congestion signaling
–Source detects (if it is capable) congestion from transmission delays
and discarded packets and reduces flow (excess delay and pkt
discard are implicit evidences of network congestion)
–Congestion control on this basis is the responsibility of end systems
and does not require action on the part of the network nodes.

12. Chapter 10 Congestion Control in Data Networks and Internets
12
Explicit congestion signaling
For explicit congestion avoidance, the network alerts end systems to
growing congestion within the network and the end systems take
steps to reduce the offered load to the network.
Direction
–Backward
–Forward
Categories
–Binary –a bit is set in a data pkt as it is forwarded by the
congested node.
–Credit-based –credits indicate how many octets or how many
pkts the source may transmit.
–rate-based –the source may transmit data a rate up to the set
limit.

13. Chapter 10 Congestion Control in Data Networks and Internets
13
Traffic Management
Considerations to refine the application of congestion
control techniques and discard policy.
Fairness
–Last-in-first-discarded may not be fair. The queues
with the highest traffic load suffer discard (fairness)
Quality of Service
–Voice, video: delay sensitive, loss insensitive
–File transfer, mail: delay insensitive, loss sensitive
–Interactive computing: delay and loss sensitive
Reservations –Contract based as in ATM networks.
–Policing: excess traffic discarded or handled on best-
effort basis

14. Chapter 10 Congestion Control in Data Networks and Internets
14
Figure 10.5

15. Chapter 10 Congestion Control in Data Networks and Internets
15
Frame Relay Congestion Control
Objectives of congestion control in FR
–Minimize frame size
–Maintain QoS
–Minimize monopolization of network
–Simple to implement, little overhead
–Minimal additional network traffic
–Resources distributed fairly
–Limit spread of congestion
–Operate effectively regardless of flow
–Have minimum impact other systems in network
–Minimize variance in QoS

16. Chapter 10 Congestion Control in Data Networks and Internets
16
Table 10.1

17. Chapter 10 Congestion Control in Data Networks and Internets
17
Traffic Rate Management
Committed Information Rate (CIR)
–Rate that network agrees to support
Aggregate of CIRs < capacity
–For node and user-network interface (access)
Committed Burst Size
–Maximum data over one interval agreed to by network
Excess Burst Size
–Maximum data over one interval that network will
attempt

18. Chapter 10 Congestion Control in Data Networks and Internets
18
Figure 10.6

19. Chapter 10 Congestion Control in Data Networks and Internets
19
Figure 10.7

20. Chapter 10 Congestion Control in Data Networks and Internets
20
Congestion Avoidance with Explicit
Signaling
2 strategies
Congestion always occurred slowly,
almost always at egress nodes
–forward explicit congestion avoidance
Congestion grew very quickly in internal
nodes and required quick action
–backward explicit congestion avoidance

21. Chapter 10 Congestion Control in Data Networks and Internets
21
2 Bits for Explicit Signaling
Forward Explicit Congestion Notification
–For traffic in same direction as received frame
–This frame has encountered congestion
Backward Explicit Congestion Notification
–For traffic in opposite direction of received
frame
–Frames transmitted may encounter congestion