, NotesNeo
Unit 4 : Congestion, QoS, LAN and Security
Syllabus :
Congestion Control, Quality of Service, QoS Improving techniques.
LAN: Ethernet, Token Bus, Token Ring, MAN Architecture- DQDB, WAN
Architectures- Frame Relay, ATM, SONET/SDH.
Network Security: Firewalls, security goals, types of attack, Introduction to
cryptography, Types of ciphers: symmetric and asymmetric key ciphers.
Section 1 : Congestion and QoS
1.1 Congestion Control
Definition:
Congestion control refers to the techniques and mechanisms used for managing
data flow to prevent network overload; ensuring efficient data transmission.
Objectives:
● Maintain Network Performance: Ensure data flows smoothly without
excessive delays.
● Prevent Packet Loss: Avoid packet drops due to overflowing buffers.
● Optimize Resource Utilization: Efficiently use network bandwidth and
resources.
● Fairness: Ensure fair bandwidth distribution among users.
Causes of Congestion:
1. High Data Load: Excessive data packets transmitted simultaneously can
exceed the network’s capacity.
2. Inefficient Routing: Suboptimal routing decisions can lead to bottlenecks
where certain paths become overloaded while others are underutilized.
3. Limited Bandwidth: Insufficient bandwidth to handle the volume of data
traffic, particularly during peak usage times.
4. Burst Traffic: Sudden spikes in traffic, such as video streaming or large file
transfers, can overwhelm the network.
5. Network Failures: Failures in network components like routers or links can
reroute traffic to other parts of the network, causing congestion.
6. Poorly Designed Network: Inadequate network design or configuration can
lead to inefficient use of resources and potential congestion points.
1
, NotesNeo
Congestion Control Techniques:
A. Open Loop (Preventive) Methods:
➢ Aim to prevent congestion before it occurs through proactive policies.
1. Traffic Shaping:
Purpose: To smooth out traffic flows by controlling the flow of bursts of data so it
doesn't overwhelm the network, maintaining a consistent transmission rate.
Techniques:
I. Leaky Bucket Algorithm:
● Concept: Imagines a bucket with a small hole at the bottom. Data goes into
the bucket and leaks out at a steady rate.
● Mechanism: Data packets are added to the bucket. If the bucket (buffer)
overflows, packets are discarded. Packets leave the bucket at a constant
rate, regardless of incoming rate.
● Pros: Keeps the data flow steady and predictable. Simplifies traffic flow to a
steady stream, effective for applications needing consistent bandwidth.
● Cons: May not fully utilize available bandwidth during low traffic periods.
II. Token Bucket Algorithm:
● Concept: Imagines a bucket that holds tokens, which are generated at a
steady rate.
● Mechanism: To send a packet, a token is required. If there are enough
tokens, packets are sent, allowing for bursts while controlling the average
rate.
● Pros: Good for allowing bursts of data while controlling the overall rate, more
flexible than leaky bucket.
● Cons: Complexity in managing token generation and consumption.
Traffic Shaping (Token Bucket Example):
● Scenario: A video streaming service that experiences bursts of high traffic.
● Implementation: Tokens are generated at a rate of 100 tokens per second.
Each video packet requires one token. When a burst of video packets arrives,
they are transmitted as long as there are tokens available. If tokens run out,
packets are queued until more tokens are generated.
● Outcome: Smooths out bursts, ensuring that average bandwidth usage stays
within limits while accommodating temporary surges.
2. Admission Control:
Purpose: To prevent too much data from entering the network at once.
Mechanism:
● The network checks current load and resources before admitting new flows.
● If resources are insufficient or network is too busy, new data flows are
rejected or delayed until the network can handle them.
● Ensures that existing traffic is not degraded by new flows.
2