In this blog, we will begin to explore the topic of queuing in QoS. The words queue and queuing may be foreign to you at this time because we do not use this terminology in everyday life in the United States. A queue is synonymous to a line. A queue (or line) normally forms because the servicing mechanism is congested. If you go to the bank and wait on line (queue), it’s because there are no available tellers. In data networks, queues form because the output interface is congested (12Mbps cannot fit on a 10Mbps interface. Instead of dropping traffic, routers and switches will queue packets into an output queue and all new packets go through the queuing system.
Let’s take the line concept and start to talk about different scheduling algorithms. The default scheduling algorithm used on most interfaces of Cisco routers and switches if FIFO (first in, first out). To understand FIFO processing, let’s re-visit the bank line analogy from the last paragraph. If there are three tellers at the bank with one line of people waiting to be serviced, the line is scheduled on a FIFO basis. FIFO queuing is problematic because there is no service differentiation. The last packet in the queue may be the most important packet in the queue, but it’s still the last packet to be serviced out of the queue. If John Chambers (CEO of Cisco) was the last person in line at the bank using a FIFO scheduling algorithm, John would be the last customer serviced. An advanced queuing algorithm would be able to identify John Chambers as a very high priority customer so John would receive immediate attention.
Weighted fair queuing (WFQ) is a queuing algorithm that has historically been enabled on E-1 (2.048Mbps), T-1 (1.544Mbps) and lower bandwidths (fractional E-1 and T-1 interfaces). The default queuing mechanism in newer IOS versions is FIFO, not WFQ as it was traditionally. WFQ only scales to interface speeds up to E-1 interfaces, with the exception of ATM interfaces. ATM interfaces support WFQ at much higher data speeds than E-1. We will investigate WFQ because the concepts in WFQ are used in CB-WFQ (class-based weighted fair queuing) and LLQ (low latency queuing). WFQ uses an automatic classification mechanism which is based on the following 6 parameters:
• Source IP address
• Destination IP address
• Source port
• Destination port
• Type of service
WFQ uses a mathematical algorithm that prioritizes interactive traffic with high IP Precedence values. The algorithm identifies interactive traffic by focusing on the packet length of the packet. Voice over IP, Citrix packets, Telnet, SSH, Instant Messaging, and remote control applications are examples of applications that are interactive in nature.
The next blog will continue our queuing conversation with coverage of class-based weighted fair queuing (CB-WFQ) and low latency queuing (LLQ).
Dennis Hartmann, CCIE No. 15651, is a consultant with www.highpoint.com and author of Implementing Cisco Unified Communications Manager, Part 1. Dennis is also a lead instructor at Global Knowledge. Dennis has various certifications, including the Cisco CCVP, CCSI, CCNP, CCIP, and the Microsoft MCSE. Dennis has various specializations including unified communications, data center, routing & switching, service provider (MPLS and optical). Dennis has worked for various Fortune 500 companies, including AT&T, Sprint, Merrill Lynch, KPMG, and Cabletron Systems. He lives with his wife and children in Hopewell Junction, New York.