You are on page 1of 6

Submitted By: Group#2

● Areej Fatima( 2125165008)


● Haiqa Talib( 2125165023)
● Malaika Afzal( 2125165032)
● Urwah Rafiq( 2125165055)
● Zaryab Iftikhar( 2125165060)

Class: BS-SE-V, Section-A

Course: Maj/SE-308 Computer Networks

Topic: Congestion, its Detection Methods, its Removal Techniques

Submitted To: Ms. Bisma Islam


Congestion:

Network congestion refers to the condition where a network experiences too much traffic, leading to
slow communication, data loss, and latency. Here are some possible examples of network
congestion:

- When many people try to access a website at the same time, the server may not be able to
handle the requests, resulting in a slow or unresponsive webpage.

- In a busy office with limited bandwidth, multiple users streaming videos or downloading large
files can cause a bottleneck that affects other users' activities, such as email or web browsing.

Causes of Congestion:

1. High data traffic: When a lot of data is being transmitted simultaneously, it can cause network
congestion, causing slowdown and packet loss.

2. Limited network capacity: If the network infrastructure doesn't have the capacity to handle the
data traffic, it can cause congestion.

3. Inefficient network design: Poor network design, including inefficient routing protocols, can
cause congestion.

4. Network equipment failure: A failing router or switch can cause congestion by reducing the
available network capacity and causing packet loss.

5. Security attacks: Denial-of-service (DoS) attacks can flood a network with traffic, causing
congestion and effectively shutting down the network.

Congestion Removal Policies in Closed-loop Congestion Control:

1. Backpressure: A congested node stops the receipt of packets from an upstream node.

Example: In a network of routers, if Router A becomes congested, it sends a back-pressure signal to


Router B, which stops sending packets to Router A until the congestion is relieved.
2. Choke packet: A router sends a choke packet to the source as an implicit signalling
mechanism.

Example: When a router becomes congested, it sends a choke packet to the source node, indicating
that the network is congested and the source should reduce its transmission rate.

3. Implicit signalling: Congestion is signalled by other actions, such as packet loss or increased
delay.

Example: When a network becomes congested, packets may be dropped or delayed, and this
implicitly signals to the source nodes that congestion is occurring. The source nodes can then
respond by reducing their transmission rate.

4. Explicit signalling: Direct communication between the congested node and the source to
signal congestion.

Example: A congested router directly sends a signal or message to the source node, explicitly
indicating that congestion is happening and requesting the source to reduce its transmission rate.

5. Load shedding: Discarding packets to reduce the load on the network.

Example: When a network becomes congested, the routers may start discarding certain packets to
alleviate congestion. For example, in a video streaming application, low-priority packets like
advertisements or non-essential data can be discarded during congestion.
Congestion Detection Techniques:

1. Packet Loss:

Detects congestion by monitoring the percentage of lost data packets.

Example: If 5 out of 100 packets fail to reach their destination, it indicates congestion.

2. Delay:

Measures the time taken for data to travel from sender to receiver.

Example: Increased delay in loading a webpage may suggest network congestion.

3. Jitter:

Examines variations in packet arrival times.

Example: VoIP calls with inconsistent audio delays due to jitter.

4. Queue Length:

Monitors the number of packets waiting in network queues.

Example: If a router queue is consistently full, congestion is likely.

5. Utilization:

Analyzes the percentage of available bandwidth in use.

Example: High utilization rates during peak hours signal potential congestion.

6. SNMP Traps:

Uses Simple Network Management Protocol traps to notify about network events.

Example: SNMP trap alerting administrators to increased network errors.

7. Flow-Based Analysis:

Examines the characteristics and patterns of data flows.

Example: Identifying a sudden surge in traffic from a specific application signaling potential
congestion.
Congestion Removal Techniques:

1. Traffic Monitoring:

● Procedure: Regularly observe and analyze the network's data flow patterns using monitoring
tools.
● Example: Using software to track the volume of data during different times of the day.
● Advantage: Helps in planning for network upgrades based on usage trends.
● Disadvantage: Reactive approach; doesn't prevent congestion in real-time.

2. Queue Length Monitoring:

● Procedure: Monitor the number of packets waiting in network queues using network
management tools.
● Example: Checking the queue length on a router during periods of heavy traffic.
● Advantage: Provides direct insight into the level of network congestion.
● Disadvantage: May not address congestion proactively; only reacts to existing issues.

3. Packet Dropping:

● Procedure: Deliberately discard packets during congestion to manage network load.


● Example: A router dropping packets to ease congestion and prevent network collapse.
● Advantage: Prevents network collapse by prioritizing essential data.
● Disadvantage: Can impact application performance; may lead to retransmissions.

4. Round-Trip Time (RTT) Monitoring:

● Procedure: Continuously monitor the time taken for a packet to travel to its destination and
back.
● Example: Observing increased RTT as an early sign of potential congestion.
● Advantage: Provides real-time insights into network performance.
● Disadvantage: Reactive; congestion might already be affecting users by the time it's detected.
5. Congestion Control Algorithms:

● Procedure: Implement automated algorithms that adjust data flow to avoid congestion.
● Example: TCP's congestion control algorithms like Additive Increase Multiplicative Decrease
(AIMD).
● Advantage: Proactively manages network resources to prevent congestion.
● Disadvantage: Implementation may be complex; requires coordination across network
components.

6. AI-Based Techniques:

● Procedure: Employ machine learning models to predict and manage congestion based on
historical data.
● Example: Using AI to forecast congestion patterns and optimize network resources.
● Advantage: Adaptive and predictive, learns from network behavior.
● Disadvantage: Implementation complexity and resource-intensive; may require extensive
training data.

7. Active Probing:

● Procedure: Actively send test packets to assess network conditions and response times.
● Example: Periodically pinging network elements to measure response times.
● Advantage: Proactively identifies potential congestion and assesses real-time conditions.
● Disadvantage: Adds some overhead to the network; may not be suitable for continuous
monitoring.

8. Flow-Based Analysis:

● Procedure: Examine the characteristics and patterns of data flows to identify congestion
sources.
● Example: Analyzing specific applications causing congestion during peak usage.
● Advantage: Pinpoints the source of congestion for targeted solutions.
● Disadvantage: May require sophisticated tools; can be complex to implement.

You might also like