CCDE 400-007: Which QoS strategy can be used avoid TCP-based app tail drops congestion?

Table of Contents

Question
Answer
Explanation 8
Reference

Question

You are designing the QoS policy for a company that is running many TCP-based applications. The company is experiencing tail drops for these applications. The company wants to use a congestion avoidance technique for these applications. Which QoS strategy can be used to fulfill the requirement?

A. weighted fair queuing
B. weighted random early detection
C. first-in first-out
D. low-latency queuing

Answer

B. weighted random early detection

Explanation 8

The correct answer to the question is B. weighted random early detection (WRED).

In a scenario where a company is running TCP-based applications and experiencing tail drops, a congestion avoidance technique is required to address this issue. Tail drops occur when the network buffers become full, resulting in packets being dropped. This can lead to reduced application performance and user experience.

Weighted random early detection (WRED) is a QoS strategy that can be used to fulfill the requirement of congestion avoidance. WRED operates at the network layer and is typically implemented on routers or switches. It helps prevent network congestion by selectively dropping packets before the network buffers become completely full.

Here’s how WRED works:

Traffic Classification: WRED classifies packets into different traffic flows based on their IP addresses, source/destination ports, or other defined criteria. In this case, the TCP-based applications would be identified as a specific traffic flow.
Setting Queue Thresholds: WRED sets thresholds for each traffic flow’s queue. These thresholds determine the point at which packets will start to be selectively dropped. The thresholds are usually based on the average queue length or congestion indicators, such as the level of buffer occupancy or average queue delay.
Random Early Drop Probability: WRED uses a random mechanism to determine which packets to drop when the queue length exceeds the configured thresholds. Instead of dropping all packets once the queue is full, WRED selectively drops packets with a probability that increases as the queue length approaches the configured maximum. This approach allows the network to provide an indication of congestion to the sender without overwhelming it with dropped packets.
TCP-Friendly Behavior: WRED is particularly effective for TCP-based applications because it provides a congestion indication to the TCP sender. When WRED starts dropping packets, TCP interprets this as an indication of network congestion and adapts by reducing its sending rate. This helps TCP-based applications to share the available bandwidth fairly and mitigate the occurrence of tail drops.

Compared to other QoS strategies listed in the options:

Weighted fair queuing (A) is a scheduling algorithm that provides fairness among different flows but does not specifically address congestion avoidance.
First-in first-out (C) is a basic queuing strategy that does not implement congestion avoidance mechanisms.
Low-latency queuing (D) is a queuing strategy that ensures low latency for high-priority traffic but does not directly address congestion avoidance.

In summary, to fulfill the requirement of implementing a congestion avoidance technique for TCP-based applications experiencing tail drops, the most appropriate QoS strategy is weighted random early detection (WRED). WRED selectively drops packets based on configured thresholds and provides a congestion indication to TCP senders, enabling TCP-friendly behavior and helping to alleviate network congestion.

Reference

Cisco Certified Design Expert CCDE v3.0 400-007 certification exam practice question and answer (Q&A) dump with detail explanation and reference available free, helpful to pass the Cisco Certified Design Expert CCDE v3.0 400-007 exam and earn Cisco Certified Design Expert CCDE v3.0 400-007 certification.