As detailed in a previous tip regarding the congestion management component of QoS, basic queuing methods (Priority, Custom, and Weighted Fair) provide the foundation behind prioritizing traffic based on their classification or characteristics. Due to the special network performance requirements essential to transmitting voice traffic across a data infrastructure, queuing techniques carve out highly available, low delay bandwidth across the data network. Because these basic queuing methods do not meet the congestion management requirements for Voice over IP traffic, hybrid queuing methods were developed to satisfy those necessities. This tip provides the details of the hybrid queuing methods available today.
Class Based Weighted Fair Queuing (CBWFQ). CBWFQ addresses the possibility of "crunching" high priority traffic using weighted fair queuing (WFQ). This class-based method classifies the traffic flows using access lists, their DSCP values, or source/destination IP addresses. Subsequently, CBWFQ assigns the appropriate weight, bandwidth, and packet limit to the class. Therefore, it can assign the necessary bandwidth to the voice "class" and queues it ahead of the other classes of traffic. If a "default" queue is configured, all unclassified traffic is assigned to that queue. Otherwise, that traffic is treated as "best effort". CBWFQ provides a method of granularly allocated bandwidth across configured classes (maximum of 64). If bandwidth is not
IP RTP Priority Queuing (PQ/WFQ). IP RTP Priority queuing incorporates the priority queuing features with the weighted fair queuing scheme. This allows the voice and other delay sensitive traffic to be placed into a priority queue in order to be serviced and sent before all other traffic. Packets are assigned to this priority queue by their UDP ports. The remaining traffic is dispatched via the WFQ scheme.
Low Latency Queuing (LLQ). LLQ is also known as the combination of priority queuing and class based weighted fair queuing or PQ-CBWFQ. Like IP RTP Priority queuing, voice and other delay sensitive traffic are placed into a priority queue in order to be serviced and sent before all other traffic. Using LLQ, packets can be assigned to this priority queue using their previously mentioned classification techniques. Therefore, it is not limited to assigning UDP ports. Another voice advantage over IP RTP Priority is that LLQ prioritizes even (call) and odd (control) voice ports whereas IP RTP can only prioritize even voice ports within the specified UDP range.
It has been shown that LLQ best meets the congestion management requirements for Voice over IP traffic. It has a priority queue for delay sensitive traffic, it is not limited to classifying traffic via UDP ports, and it provides guaranteed bandwidth and latency.
The next tip will delve into the congestion avoidance techniques found within the Quality of Service Model.
Richard Parsons (CCIE#5719) is a Principal Architect for SBC Communications with a focus on network planning, design, and implementation. He has built a solid foundation in networking concepts, advanced troubleshooting, and monitoring in areas such as optical, ATM, VoIP, routed, routing, and storage infrastructures. Rich resides in Atlanta GA, and is a graduate of Clemson University. His background includes consulting positions at International Network Services, Lucent, and Callisma.
This was first published in May 2004