/mcpQuality of Service for Voice over IP QoS Queueing Mechanisms 9 QoSVoIP.mif Other QoS Queueing Mechanisms Several other queueing methods are available. For example, Modified Deficit Round Robin (MDRR) is a queueing mechanism available on the Cisco 12000 series Gigabit Switch Routers (GSRs) that allows bandwidth guarantees and priority servicing based on IP Precedence, DSCP, and MPLS EXP classes. MDRR supports one priority queue, seven reserved queues, and one multicast queue. Once again, VoIP requires priority but there are data applications that cannot be starved and need bandwidth guarantees. You can use any queueing mechanism that effectively gives VoIP high priority, but we recommend LLQ. Table 1 describes some of the available software queueing mechanisms. Table 1 Software Queueing Mechanisms Software Queueing Mechanism Description Benefits Limitations FIFO Packets arrive and leave the queue in exactly the same order. Simple configuration and fast operation. No priority servicing or bandwidth guarantees are possible. WFQ A hashing algorithm places flows into separate queues where weights are used to determine how many packets are serviced at a time. You define weights by setting IP Precedence and DSCP values. Simple configuration. Default on links less than 2 Mbps. No priority servicing or bandwidth guarantees are possible. Custom Queueing (CQ) Traffic is classified into multiple queues with configurable queue limits. The queue limits are calculated based on average packet size, maximum transmission unit (MTU), and the percentage of bandwidth to be allocated. Queue limits (in number of bytes) are dequeued for each queue, therefore providing the allocated bandwidth statistically. Has been available for a few years and allows approximate bandwidth allocation for different queues. No priority servicing is possible. Bandwidth guarantees are approximate, and there are a limited number of queues. Configuration is relatively difficult. Priority Queueing (PQ) Traffic is classified into high, medium, normal, and low priority queues. The high priority traffic is serviced first, then medium priority traffic, followed by normal and low priority traffic. Has been available for a few years and provides priority servicing. Higher priority traffic can starve the lower priority queues of bandwidth. No bandwidth guarantees are possible.