Catalyst 4500 Series Switch Software Configuration Guide, 12.2(54)SG

Prioritization

QoS implementation is based on the DiffServ architecture. This architecture specifies that each packet is classified upon entry into the network. The classification is carried in the IP packet header, using 6 bits from the deprecated IP type of service (ToS) field to carry the classification ( class) information. Classification can also be carried in the Layer 2 frame. These special bits in the Layer 2 frame or a Layer 3 packet are described here and shown in Figure 37-1:

• Prioritization values in Layer 2 frames:

– Layer 2 Inter-Switch Link (ISL) frame headers have a 1-byte User field that carries an IEEE 802.1p class of service (CoS) value in the three least-significant bits. On interfaces configured as Layer 2 ISL trunks, all traffic is in ISL frames.

– Layer 2 802.1Q frame headers have a 2-byte Tag Control Information field that carries the CoS value in the three most-significant bits, which are called the User Priority bits. On interfaces configured as Layer 2 802.1Q trunks, all traffic is in 802.1Q frames except for traffic in the native VLAN.

– Other frame types cannot carry Layer 2 CoS values.

– Layer 2 CoS values range from 0 for low priority to 7 for high priority.

• Prioritization bits in Layer 3 packets:

– Layer 3 IP packets can carry either an IP precedence value or a Differentiated Services Code Point (DSCP) value. QoS supports the use of either value because DSCP values are backward-compatible with IP precedence values.

– IP precedence values range from 0 to 7.

– DSCP values range from 0 to 63.

Figure 37-1 QoS Classification Layers in Frames and Packets

All switches and routers across the Internet rely on the class information to provide the same forwarding treatment to packets with the same class information and different treatment to packets with different class information. The class information in the packet can be assigned by end hosts or by switches or routers along the way, based on a configured policy, detailed examination of the packet, or both. Detailed examination of the packet is expected to happen closer to the edge of the network so that the core switches and routers are not overloaded.

Switches and routers along the path can use the class information to limit the amount of resources allocated per traffic class. The behavior of an individual device when handling traffic in the DiffServ architecture is called per-hop behavior. If all devices along a path provide a consistent per-hop behavior, you can construct an end-to-end QoS solution.

Implementing QoS in your network can be a simple or complex task and depends on the QoS features offered by your internetworking devices, the traffic types and patterns in your network, and the granularity of control you need over incoming and outgoing traffic.

QoS Terminology

The following terms are used when discussing QoS features:

• Packets carry traffic at Layer 3.
• Frames carry traffic at Layer 2. Layer 2 frames carry Layer 3 packets.
• Labels are prioritization values carried in Layer 3 packets and Layer 2 frames:

– Layer 2 class of service (CoS) values, which range between zero for low priority and seven for high priority:

Layer 2 Inter-Switch Link (ISL) frame headers have a 1-byte User field that carries an IEEE 802.1p CoS value in the three least significant bits.

Layer 2 802.1Q frame headers have a 2-byte Tag Control Information field that carries the CoS value in the three most significant bits, which are called the User Priority bits.

Other frame types cannot carry Layer 2 CoS values.

Note On interfaces configured as Layer 2 ISL trunks, all traffic is in ISL frames. On interfaces configured as Layer 2 802.1Q trunks, all traffic is in 802.1Q frames except for traffic in the native VLAN.

– Layer 3 IP precedence values—The IP version 4 specification defines the three most significant bits of the 1-byte ToS field as IP precedence. IP precedence values range between zero for low priority and seven for high priority.

– Layer 3 differentiated services code point (DSCP) values—The Internet Engineering Task Force (IETF) has defined the six most significant bits of the 1-byte IP ToS field as the DSCP. The per-hop behavior represented by a particular DSCP value is configurable. DSCP values range between 0 and 63. For more information, see the “Configuring DSCP Maps” section.

Note Layer 3 IP packets can carry either an IP precedence value or a DSCP value. QoS supports the use of either value, since DSCP values are backwards compatible with IP precedence values. See Table 37-1.

• Classification is the selection of traffic to be marked.
• Marking , according to RFC 2475, is the process of setting a Layer 3 DSCP value in a packet; in this publication, the definition of marking is extended to include setting Layer 2 CoS values.
• Scheduling is the assignment of Layer 2 frames to a queue. QoS assigns frames to a queue based on internal DSCP values as shown in Internal DSCP Values.
• Policing is limiting bandwidth used by a flow of traffic. Policing can mark or drop traffic.

Basic QoS Model

Figure 37-2 shows the basic QoS model (also referred to as Switch QoS model; it is not MQC compliant). Actions at the ingress and egress interfaces include classifying traffic, policing, and marking:

• Classifying distinguishes one kind of traffic from another. The process generates an internal DSCP for a packet, which identifies all the future QoS actions to be performed on this packet. For more information, see the “Classification” section.
• Policing determines whether a packet is in or out of profile by comparing the traffic rate to the configured policer, which limits the bandwidth consumed by a flow of traffic. The result of this determination is passed to the marker. For more information, see the “Policing and Marking” section.
• Marking evaluates the policer configuration information regarding the action to be taken when a packet is out of profile and decides what to do with the packet (pass through a packet without modification, mark down the DSCP value in the packet, or drop the packet). For more information, see the “Policing and Marking” section.

Actions at the egress interface include queueing and scheduling:

• Queueing evaluates the internal DSCP and determines which of the four egress queues in which to place the packet.
• Scheduling services the four egress (transmit) queues based on the sharing and shaping configuration of the egress (transmit) port. Sharing and shaping configurations are described in the “Queueing and Scheduling” section.

Figure 37-2 Basic QoS Model

Classification

Classification is the process of distinguishing one kind of traffic from another by examining the fields in the packet. Classification is enabled only if QoS is globally enabled on the switch. By default, QoS is globally disabled, so no classification occurs.

You specify which fields in the frame or packet that you want to use to classify incoming traffic.

Classification options are shown in Figure 37-3.

For non-IP traffic, you have the following classification options:

• Use the port default. If the packet is a non-IP packet, assign the default port DSCP value to the incoming packet.
• Trust the CoS value in the incoming frame (configure the port to trust CoS). Next, use the configurable CoS-to-DSCP map to generate the internal DSCP value. Layer 2 ISL frame headers carry the CoS value in the three least-significant bits of the 1-byte User field. Layer 2 802.1Q frame headers carry the CoS value in the three most-significant bits of the Tag Control Information field. CoS values range from 0 for low priority to 7 for high priority. If the frame does not contain a CoS value, assign the default port CoS to the incoming frame.

The trust DSCP configuration is meaningless for non-IP traffic. If you configure a port with trust DSCP and non-IP traffic is received, the switch assigns the default port DSCP.

Note On a Catalyst 4948-10GE, Supervisor V-10GE, and Supervisor V, when you send non-IP traffic (such as IPX) from port 1 with.1Q tag and a Pri=x value, for all x values, the transmit CoS at the output interface varies as srcMac changes.

"trust dscp" in a policy map only works for IP packets. If a non-IP packet is matched by class with the “trust dscp" action, it could be transmitted with a random CoS value rather than treating the "trust dscp" as no-OP as assumed otherwise.

Workaround: Make the classification criteria in you class-maps more granular so that "trust dscp" is applied to a class that won't match non-IP packet. Separate the class-map matching both IP and non-IP packets into a set of class-maps so that "trust dscp" is applied to a class that matches only IPv4 traffic. For class-maps match non-IPv4 traffic, "trust cos" can be used.

For IP traffic, you have the following classification options:

• Trust the IP DSCP in the incoming packet (configure the port to trust DSCP), and assign the same DSCP to the packet for internal use. The IETF defines the six most-significant bits of the 1-byte type of service (ToS) field as the DSCP. The priority represented by a particular DSCP value is configurable. DSCP values range from 0 to 63.
• Trust the CoS value (if present) in the incoming packet, and generate the DSCP by using the CoS-to-DSCP map.
• Perform the classification based on a configured IP standard or extended ACL, which examines various fields in the IP header. If no ACL is configured, the packet is assigned the default DSCP based on the trust state of the ingress port; otherwise, the policy map specifies the DSCP to assign to the incoming frame.

Note It is not possible to classify traffic based on the markings performed by an input QoS policy. In the Catalyst 4500 platform, the input and output QoS lookup happen in parallel, and therefore, input marked DSCP value cannot be used to classify traffic in the output QoS policy.

Note It is not possible to classify traffic based on internal DSCP. The internal DSCP is purely an internal classification mechanism used for all packets to determine transmit queue and transmit CoS values only.

For information on the maps described in this section, see the “Mapping Tables” section. For configuration information on port trust states, see the “Configuring the Trust State of Interfaces” section.

Figure 37-3 Classification Flowchart

Classification Based on QoS ACLs

A packet can be classified for QoS using multiple match criteria, and the classification can specify whether the packet should match all of the specified match criteria or at least one of the match criteria. To define a QoS classifier, you can provide the match criteria using the match statements in a class map. In the match statements, you can specify the fields in the packet to match on, or you can use IP standard or IP extended ACLs. For more information, see the “Classification Based on Class Maps and Policy Maps” section.

If the class map is configured to match all the match criteria, then a packet must satisfy all the match statements in the class map before the QoS action is taken. The QoS action for the packet is not taken if the packet does not match even one match criterion in the class map.

If the class map is configured to match at least one match criterion, then a packet must satisfy at least one of the match statements in the class map before the QoS action is taken. The QoS action for the packet is not taken if the packet does not match any match criteria in the class map.

Note When you use the IP standard and IP extended ACLs, the permit and deny ACEs in the ACL have a slightly different meaning in the QoS context.

• If a packet encounters (and satisfies) an ACE with a permit, then the packet matches the match criterion in the QoS classification.
• If a packet encounters (and satisfies) an ACE with a deny, then the packet does not match the match criterion in the QoS classification.
• If no match with a permit action is encountered and all the ACEs have been examined, then the packet does not match the criterion in the QoS classification.

Note When creating an access list, remember that, by default, the end of the access list contains an implicit deny statement for everything if it did not find a match before reaching the end.

After a traffic class has been defined with the class map, you can create a policy that defines the QoS actions for a traffic class. A policy might contain multiple classes with actions specified for each one of them. A policy might include commands to classify the class as a particular aggregate (for example, assign a DSCP) or rate-limit the class. This policy is then attached to a particular port on which it becomes effective.

You implement IP ACLs to classify IP traffic by using the access-list global configuration command. For configuration information, see the “Configuring a QoS Policy” section.

Classification Based on Class Maps and Policy Maps

A class map is a mechanism that you use to isolate and name a specific traffic flow (or class) from all other traffic. The class map defines the criterion used to match against a specific traffic flow to further classify it; the criteria can include matching the access group defined by the ACL or matching a specific list of DSCP, IP precedence, or Layer 2 CoS values. If you have more than one type of traffic that you want to classify, you can create another class map and use a different name. After a packet is matched against the class map criteria, you can specify the QoS actions by using a policy map.

A policy map specifies the QoS actions for the traffic classes. Actions can include trusting the CoS or DSCP values in the traffic class; setting a specific DSCP or IP precedence value in the traffic class; or specifying the traffic bandwidth limitations and the action to take when the traffic is out of profile. Before a policy map can be effective, you must attach it to an interface.

You create a class map by using the class-map global configuration command. When you enter the class-map command, the switch enters the class map configuration mode. In this mode, you define the match criteria for the traffic by using the match class-map configuration command.

You create and name a policy map by using the policy-map global configuration command. When you enter this command, the switch enters the policy-map configuration mode. In this mode, you specify the actions to take on a specific traffic class by using the trust or set policy map configuration and policy map class configuration commands. To make the policy map effective, you attach it to an interface by using the service-policy interface configuration command.

The policy map can also contain commands that define the policer, (the bandwidth limitations of the traffic) and the action to take if the limits are exceeded. For more information, see the “Policing and Marking” section.

A policy map also has these characteristics:

• A policy map can contain up to 255 class statements.
• You can have different classes within a policy map.
• A policy map trust state supersedes an interface trust state.

For configuration information, see the “Configuring a QoS Policy” section.

Policing and Marking

After a packet is classified and has an internal DSCP value assigned to it, the policing and marking process can begin as shown in Figure 37-4.

Policing involves creating a policer that specifies the bandwidth limits for the traffic. Packets that exceed the limits are out of profile or nonconforming . Each policer specifies the action to take for packets that are in or out of profile. These actions, carried out by the marker, include passing using the packet without modification, dropping the packet, or marking down the packet with a new DSCP value that is obtained from the configurable policed-DSCP map. For information on the policed-DSCP map, see the “Mapping Tables” section.

You can create these types of policers:

• Individual

QoS applies the bandwidth limits specified in the policer separately to each matched traffic class for each port/VLAN to which the policy map is attached to. You configure this type of policer within a policy map by using the police command under policy map class configuration mode.

• Aggregate

QoS applies the bandwidth limits specified in an aggregate policer cumulatively to all matched traffic flows. You configure this type of policer by specifying the aggregate policer name within a policy map by using the police aggregate policy-map configuration command. You specify the bandwidth limits of the policer by using the qos aggregate-policer global configuration command. In this way, the aggregate policer is shared by multiple classes of traffic within a policy map.

• Flow or microflow

With flow-based policing, all the identified flows are policed to the specified rate individually. Because the flows are dynamic, key distinguishing fields must be configured in class maps. Two flow-matching options are provided: source ip based (each flow with unique source IP address is considered a new flow) and destination ip based (each flow with unique destination IP address is considered as new flow). For information on flow-based policer configuration, see “Configuring User-Based Rate-Limiting” section.

When configuring policing and policers, observe these guidelines:

• On Supervisor Engine 6-E, Supervisor Engine 6L-E, Catalyst 4900M, and Catalyst 4948E, policers account only for the Layer 2 header length when calculating policer rates. In contrast, shapers account for header length as well as IPG in rate calculations.

By default, on Supervisor Engine V-10GE and earlier supervisor engines, policer or shaping computation involves only IP payload information. You can include Layer 2 lengths and any overhead with the qos account l2 encapsulation command.

For non-IP packets, the Layer 2 length as specified in the Layer 2 Header is used by the policer for policing computation. To specify additional Layer 2 encapsulation length when policing IP packets, use the qos account layer2 encapsulation command.

• By default, no policers are configured.
• Only the average rate and committed burst parameters are configurable.
• Policing for individual and aggregate policers can occur in ingress and egress interfaces.

– With the Supervisor Engine V-10GE (WS-X4516-10GE), 8192 policers are supported on ingress and on egress.

– With all other supervisor engines, 1024 policers are supported on ingress and on egress.

– The accuracy of the policer configured is +/- 1.5 percent.

Note Four policers in ingress and egress direction are reserved.

• Policers can be of individual or aggregate type. On the Supervisor Engine V-10GE, flow-based policers are supported.
• Policing for flow policers can occur on ingress Layer 3 interfaces only.

– 512 unique flow policers can be configured on the Supervisor Engine V-10GE.

Note Because one flow policer is reserved by software, 511 unique flow policers can be defined.

– Greater than 100,000 flows can be microflow policed.

Note Microflow currently supports two flow matching options (source IP address based and destination IP address based). When microflow policing is used together with NetFlow Statistics Collection, full flow statistics for the flows matching the source IP address or destination IP address are not available. For information on configuring NetFlow Statistics, refer to “Enabling NetFlow Statistics Collection” section.

• On an interface configured for QoS, all traffic received or sent using the interface is classified, policed, and marked according to the policy map attached to the interface. However, if the interface is configured to use VLAN-based QoS (using the qos vlan-based command), the traffic received or sent using the interface is classified, policed, and marked according to the policy map attached to the VLAN (configured on the VLAN interface) to which the packet belongs. If there is no policy map attached to the VLAN to which the packet belongs, the policy map attached to the interface is used.
• For 2 rate 3 colors (2r3c) policers, if no explicit violation-action is specified, the exceed-action is used as the violate-action.

After you configure the policy map and policing actions, attach the policy to an ingress or egress interface by using the service-policy interface configuration command. For configuration information, see the “Configuring a QoS Policy” section and the “Creating Named Aggregate Policers” section.

Figure 37-4 Policing and Marking Flowchart

Internal DSCP Values

The following sections describe the internal DSCP values:

• Internal DSCP Sources
• Egress ToS and CoS Sources

Internal DSCP Sources

During processing, QoS represents the priority of all traffic (including non-IP traffic) with an internal DSCP value. QoS derives the internal DSCP value from the following:

• For trust-CoS traffic, from received or ingress interface Layer 2 CoS values
• For trust-DSCP traffic, from received or ingress interface DSCP values
• For untrusted traffic, from ingress interface DSCP value

The trust state of traffic is the trust state of the ingress interface unless set otherwise by a policy action for this traffic class.

QoS uses configurable mapping tables to derive the internal 6-bit DSCP value from CoS, which are 3-bit values (see the“Configuring DSCP Maps” section).

Egress ToS and CoS Sources

For egress IP traffic, QoS creates a ToS byte from the internal DSCP value and sends it to the egress interface to be written into IP packets. For trust-dscp and untrusted IP traffic, the ToS byte includes the original 2 least-significant bits from the received ToS byte.

Note The internal ToS value can mimic an IP precedence value (see Table 37-1).

For all egress traffic, QoS uses a configurable mapping table to derive a CoS value from the internal ToS value associated with traffic (see the “Configuring the DSCP-to-CoS Map” section). QoS sends the CoS value to be written into ISL and 802.1Q frames.

For traffic received on an ingress interface configured to trust CoS using the qos trust cos command, the transmit CoS is always the incoming packet CoS (or the ingress interface default CoS if the packet is received untagged).

When the interface trust state is not configured to trust dscp using the qos trust dscp command, the security and QoS ACL classification always use the interface DSCP and not the incoming packet DSCP.

Mapping Tables

During QoS processing, the switch represents the priority of all traffic (including non-IP traffic) with an internal DSCP value:

• During classification, QoS uses configurable mapping tables to derive the internal DSCP (a 6-bit value) from received CoS. These maps include the CoS-to-DSCP map.
• During policing, QoS can assign another DSCP value to an IP or non-IP packet (if the packet is out of profile and the policer specifies a marked down DSCP value). This configurable map is called the policed-DSCP map.
• Before the traffic reaches the scheduling stage, QoS uses the internal DSCP to select one of the four egress queues for output processing. The DSCP-to-egress queue mapping can be configured using the qos map dscp to tx-queue command.

The CoS-to-DSCP and DSCP-to-CoS map have default values that might or might not be appropriate for your network.

For configuration information, see the “Configuring DSCP Maps” section.

Queueing and Scheduling

Each physical port has four transmit queues (egress queues). Each packet that needs to be transmitted is enqueued to one of the transmit queues. The transmit queues are then serviced based on the transmit queue scheduling algorithm.

Once the final transmit DSCP is computed (including any markdown of DSCP), the transmit DSCP to transmit queue mapping configuration determines the transmit queue. The packet is placed in the transmit queue of the transmit port, determined from the transmit DSCP. Use the qos map dscp to tx-queue command to configure the transmit DSCP to transmit queue mapping. The transmit DSCP is the internal DSCP value if the packet is a non-IP packet as determined by the QoS policies and trust configuration on the ingress and egress ports.

For configuration information, see the “Configuring Transmit Queues” section.

Active Queue Management

Active queue management (AQM) informs you about congestion before a buffer overflow occurs. AQM is done using dynamic buffer limiting (DBL). DBL tracks the queue length for each traffic flow in the switch. When the queue length of a flow exceeds its limit, DBL drop packets or set the Explicit Congestion Notification (ECN) bits in the packet headers.

DBL classifies flows in two categories, adaptive and aggressive. Adaptive flows reduce the rate of packet transmission once it receives congestion notification. Aggressive flows do not take any corrective action in response to congestion notification. For every active flow the switch maintains two parameters, “buffersUsed” and “credits.” All flows start with “max-credits,” a global parameter. When a flow with credits less than “aggressive-credits” (another global parameter) it is considered an aggressive flow and is given a small buffer limit called “aggressiveBufferLimit.”

Queue length is measured by the number of packets. The number of packets in the queue determines the amount of buffer space that a flow is given. When a flow has a high queue length the computed value is lowered. This allows new incoming flows to receive buffer space in the queue. This allows all flows to get a proportional share of packets using the queue.

Because 4 transmit queues exist per interface and DBL is a per-queue mechanism, DSCP values can make DBL application more complex.

The following table provides the default DSCP-to-transmit queue mapping:

For example, if you are sending two streams, one with a DSCP of 16 and other with a value of 0, they will transmit from different queues. Even though an aggressive flow in txQueue 2 (packets with DSCP of 16) can saturate the link, packets with a DSCP of 0 are not blocked by the aggressive flow, because they transmit from txQueue 1. Even without DBL, packets whose DSCP value places them in txQueue 1, 3, or 4 are not dropped because of the aggressive flow.

Sharing Link Bandwidth Among Transmit Queues

The four transmit queues for a transmit port share the available link bandwidth of that transmit port. You can set the link bandwidth to be shared differently among the transmit queues using bandwidth command in interface transmit queue configuration mode. With this command, you assign the minimum guaranteed bandwidth for each transmit queue.

By default, all queues are scheduled in a round-robin manner.

For systems using Supervisor Engine II-Plus, Supervisor Engine II-Plus TS, Supervisor Engine III, and Supervisor Engine IV, bandwidth can be configured on these ports only:

• Uplink ports on supervisor engines
• Ports on the WS-X4306-GB GBIC module
• Ports on the WS-X4506-GB-T CSFP module
• The two 1000BASE-X ports on the WS-X4232-GB-RJ module
• The first two ports on the WS-X4418-GB module
• The two 1000BASE-X ports on the WS-X4412-2GB-TX module

For systems using Supervisor Engine V, bandwidth can be configured on all ports (10/100 Fast Ethernet, 10/100/1000BASE-T, and 1000BASE-X).

Strict Priority / Low Latency Queueing

You can configure transmit queue 3 on each port with higher priority using the priority high tx-queue configuration command in the interface configuration mode. When transmit queue 3 is configured with higher priority, packets in transmit queue 3 are scheduled ahead of packets in other queues.

When transmit queue 3 is configured at a higher priority, the packets are scheduled for transmission before the other transmit queues only if it has not met the allocated bandwidth sharing configuration. Any traffic that exceeds the configured shape rate is queued and transmitted at the configured rate. If the burst of traffic, exceeds the size of the queue, packets are dropped to maintain transmission at the configured shape rate.

Traffic Shaping

Traffic shaping provides the ability to control the rate of outgoing traffic in order to make sure that the traffic conforms to the maximum rate of transmission contracted for it. Traffic that meets certain profile can be shaped to meet the downstream traffic rate requirements to handle any data rate mismatches.

Each transmit queue can be configured to transmit a maximum rate using the shape command. The configuration allows you to specify the maximum rate of traffic. Any traffic that exceeds the configured shape rate is queued and transmitted at the configured rate. If the burst of traffic exceeds the size of the queue, packets are dropped to maintain transmission at the configured shape rate.

Packet Modification

A packet is classified, policed, and queued to provide QoS. Packet modifications can occur during this process:

• For IP packets, classification involves assigning a DSCP to the packet. However, the packet is not modified at this stage; only an indication of the assigned DSCP is carried along. This is because QoS classification and ACL lookup occur in parallel; the ACL might specify that the packet should be denied and logged. In this situation, the packet is forwarded with its original DSCP to the CPU, where it is again processed through ACL software.
• For non-IP packets, classification involves assigning an internal DSCP to the packet, but because there is no DSCP in the non-IP packet, no overwrite occurs. Instead, the internal DSCP is used both for queueing and scheduling decisions and for writing the CoS priority value in the tag if the packet is being transmitted on either an ISL or 802.1Q trunk port.
• During policing, IP and non-IP packets can have another DSCP assigned to them (if they are out of profile and the policer specifies a markdown DSCP). Once again, the DSCP in the packet is not modified, but an indication of the marked-down value is carried along. For IP packets, the packet modification occurs at a later stage.

Note If you are running Supervisor Engines II-Plus, II-Plus-10GE, IV, V, V-10GE, or using a ME Catalyst 4900, Catalyst 4900, or Catalyst 4900-10GE series switch, you must enter the qos rewrite ip dscp command to enable update of packets and DSCP/ToS fields based on the switch’s QoS policies.

Per-Port Per-VLAN QoS

Per-port per-VLAN QoS (PVQoS) offers differentiated quality of services to individual VLANs on a trunk port. It enables service providers to rate-limit individual VLAN-based services on each trunk port to a business or a residence. In an enterprise voice-over-IP environment, it can be used to rate-limit voice VLAN even if an attacker impersonates a Cisco IP phone. A per-port per-VLAN service policy can be separately applied to either ingress or egress traffic.

QoS and Software Processed Packets

The Catalyst 4500 platform does not apply the QoS marking or policing configuration for any packets that are forwarded or generated by the Cisco IOS software. This means that any input or output QoS policy configured on the port or VLAN is not applied to packets if the Cisco IOS is forwarding or generating packets.

However, Cisco IOS marks all the generated control packets appropriately and uses the internal IP DSCP to determine the transmit queue on the output transmission interface. For IP packets, the internal IP DSCP is the IP DSCP field in the IP packet. For non-IP packets, Cisco IOS assigns a packet priority internally and maps it to an internal IP DSCP value.

Cisco IOS assigns an IP precedence of 6 to routing protocol packets on the control plane. As noted in RFC 791, “The Internetwork Control designation is intended for use by gateway control originators only.” Specifically, Cisco IOS marks the following IP-based control packets: Open Shortest Path First (OSPF), Routing Information Protocol (RIP), Enhanced Interior Gateway Routing Protocol (EIGRP) hellos, and keepalives. Telnet packets to and from the router also receive an IP precedence value of 6. The assigned value remains with the packets when the output interface transmits them into the network.

For Layer 2 control protocols, the software assigns an internal IP DSCP. Typically, Layer 2 control protocol packets are assigned an internal DSCP value of 48 (corresponding to an IP precedence value of 6).

The internal IP DSCP is used to determine the transmit queue to which the packet is enqueued on the transmission interface. See “Configuring Transmit Queues” section for details on how to configure the DSCP to transmit queues.

The internal IP DSCP is also used to determine the transmit CoS marking if the packet is transmitted with a IEEE 802.1q or ISL tag on a trunk interface. See “Configuring the DSCP-to-CoS Map” section for details on how to configure the DSCP to CoS mapping.

Default QoS Configuration

Table 37-2 shows the QoS default configuration.

Configuration Guidelines

Before beginning the QoS configuration, you should be aware of this information:

• If you have EtherChannel ports configured on your switch, you must configure QoS classification and policing on the EtherChannel. The transmit queue configuration must be configured on the individual physical ports that comprise the EtherChannel.
• If the IP fragments match the source and destination configured in the ACL used to classify the traffic for quality of service, but do not match the Layer 4 port numbers in the ACL, they are still matched with the ACL and may get prioritized. If the desired behavior is to give best-effort service to IP fragments, the following two ACEs should be added to the ACL used to classify the traffic:

• It is not possible to match IP options against configured IP extended ACLs to enforce QoS. These packets are sent to the CPU and processed by software. IP options are denoted by fields in the IP header.
• Control traffic (such as spanning tree BPDUs and routing update packets) received by the switch are subject to all ingress QoS processing.
• You cannot use set commands in policy maps if IP routing is disabled (enabled by default).
• On a dot1q tunnel port, only Layer 2 match criteria can be applied to tagged packets. However, all match criteria can be applied for untagged packets.
• On a trunk port, only Layer 2 match criteria can be applied to packets with multiple 802.1q tags.

Note QoS processes both unicast and multicast traffic.

Enabling QoS Globally

To enable QoS globally, perform this task:

This example shows how to enable QoS globally and to verify the configuration:

Enabling IP DSCP Rewrite

You must enter the qos rewrite ip dscp command to enable the update of packets CoS, DSCP and ToS fields based on the switch’s QoS policies.

To enable IP DSCP rewrite, perform this task:

This example shows how to enable IP DSCP rewrite and to verify the configuration:

Note The qos rewrite ip dscp command is not supported on Supervisor Engine 6-E, Supervisor Engine 6L-E, Catalyst 4900M, and Catalyst 4948E.

If you disable IP DSCP rewrite and enable QoS globally, the following events occur:

• The ToS byte on the IP packet is not modified.
• Marked and marked-down DSCP values are used for queueing.
• The internally derived DSCP (as per the trust configuration on the interface or VLAN policy) is used for transmit queue and Layer 2 CoS determination. The DSCP is not rewritten on the IP packet header.

If you disable QoS, the DSCP of the incoming packet are preserved and are not rewritten.

Configuring a Trusted Boundary to Ensure Port Security

In a typical network, you connect a Cisco IP phone to a switch port as discussed in Chapter38, “Configuring Voice Interfaces” Traffic sent from the telephone to the switch is typically marked with a tag that uses the 802.1Q header. The header contains the VLAN information and the class of service (CoS) 3-bit field, which determines the priority of the packet. For most Cisco IP phone configurations, the traffic sent from the telephone to the switch is trusted to ensure that voice traffic is properly prioritized over other types of traffic in the network. By using the qos trust cos interface configuration command, you can configure the switch port to which the telephone is connected to trust the CoS labels of all traffic received on that port.

Note Starting with Cisco IOS Release 12.2(31)SG, Supervisor Engine V-10GE allows you to classify traffic based on packet's IP DSCP value irrespective of the port trust state. Because of this, even when a Cisco IP phone is not detected, data traffic can be classified based on IP DSCP values. Output queue selection is not impacted by this new behavior. It is still based on the incoming port trust configuration. For information on configuring transmit queues, refer to the “Configuring Transmit Queues” section.

In some situations, you also might connect a PC or workstation to the Cisco IP phone. In this case, you can use the switchport priority extend cos interface configuration command to configure the telephone using the switch CLI to override the priority of the traffic received from the PC. With this command, you can prevent a PC from taking advantage of a high-priority data queue.

However, if a user bypasses the telephone and connects the PC directly to the switch, the CoS labels generated by the PC are trusted by the switch (because of the trusted CoS setting) and can allow misuse of high-priority queues. The trusted boundary feature solves this problem by using the CDP to detect the presence of a Cisco IP phone (such as the Cisco IP phone 7910, 7935, 7940, and 7960) on a switch port.

Note If CDP is not running on the switch globally or on the port in question, trusted boundary does not work.

When you configure trusted boundary on a port, trust is disabled. When a phone is plugged in and detected, trust is enabled. (It may take a few minutes to detect the phone.) Now, when a phone is unplugged (and not detected), the trusted boundary feature disables the trusted setting on the switch port and prevents misuse of a high-priority queue.

Note On a given port, the Cisco IP phone discovery information is not maintained on the standby supervisor engine. When the standby supervisor engine becomes active, it rediscovers the Cisco IP phone through CDP. For a short time, the port are not in the trust state after the SSO switchover.

To enable trusted boundary on a port, perform this task:

To disable the trusted boundary feature, use the no qos trust device cisco-phone interface configuration command.

Enabling Dynamic Buffer Limiting

Note Supervisor Engine 6-E, Supervisor Engine 6L-E, Catalyst 4900M, and Catalyst 4948E support DBL using the MQC CLI, but not with the older QoS DBL CLI.

Dynamic buffer limiting (DBL) provides active queue management on Catalyst 4500 platforms. Refer to “Active Queue Management” section for details.

Through selective DBL, you can select the flows that are subjected to the DBL algorithm. You can enable DBL globally on specific IP DSCP values or on specific CoS values.

This section includes the following topics:

• Enabling DBL Globally
• Selectively Enable DBL

Enabling DBL Globally

To enable DBL globally on the switch, perform this task:

This example shows how to enable DBL globally and verify the configuration:

You can enable DBL on the egress interface direction by applying a service policy:

Selectively Enable DBL

Note Selective DBL is not supported on Supervisor Engine 6-E, Supervisor Engine 6L-E, Catalyst 4900M, and Catalyst 4948E.

DSCP values enable you to selectively apply DBL for IP packets only (single or untagged). Refer to the “Enabling DBL on Specific IP DSCP Values” section.

To selectively apply DBL for non-IP packets or double-tagged packets (such as Q-in-Q), you must use CoS values. Refer to the “Enabling DBL on Specific CoS Values” section.

The following sections describe how to enable the packets:

• Enabling DBL on Specific IP DSCP Values
• Enabling DBL on Specific CoS Values

Enabling DBL on Specific IP DSCP Values

DBL action is performed on transmit queues (four per interface). You manage the mapping from IP DSCP to transmit queues with the qos map dscp dscp-values to tx-queue queue-id command. Refer to “Configuring Transmit Queues” section for details.

To enable DBL on specific IP DSCP values, perform this task:

This example shows how to selectively enable DBL on the DSCP values 1 through 10:

This example shows how to selectively disable DBL on DSCP values 1 through 10 and to verify the configuration:

Although you apply DBL based on class attributes other than DSCP, you need to attach a policy map to an egress interface. See “Configuring Policy-Map Class Actions” section.

Provided the value has been set according to your network policies, you must configure trust DSCP on the ingress interface of the aggressive flow that DBL will throttle:

Enabling DBL on Specific CoS Values

You might need to use CoS values to selectively apply DBL if you intend to use non-IP packets or double-tagged packets (for example, Q-in-Q).

For single-tagged IP packets, use the following approach. Specify the global qos dbl dscp-based command as shown in the “Enabling DBL on Specific IP DSCP Values” section.

To enable DBL non-IP or double-tagged packets, perform this task:

Note For more details on using CoS Mutation, refer to the “Configuring CoS Mutation” section.

To selectively enable DBL on CoS values 2 and 3:

Creating Named Aggregate Policers

To create a named aggregate policer, perform this task:

An aggregate policer can be applied to one or more interfaces. However, if you apply the same policer to the input direction on one interface and to the output direction on a different interface, then you have created the equivalent of two different aggregate policers in the switching engine. Each policer has the same policing parameters, with one policing the ingress traffic on one interface and the other policing the egress traffic on another interface. If an aggregate policer is applied to multiple interfaces in the same direction, then only one instance of the policer is created in the switching engine.

Similarly, an aggregate policer can be applied to a port or to a VLAN. If you apply the same aggregate policer to a port and to a VLAN, then you have created the equivalent of two different aggregate policers in the switching engine. Each policer has the same policing parameters, with one policing the traffic on the configured port and the other policing the traffic on the configured VLAN. If an aggregate policer is applied to only ports or only VLANs, then only one instance of the policer is created in the switching engine.

In effect, if you apply a single aggregate policer to ports and VLANs in different directions, then you have created the equivalent of four aggregate policers: one for all ports sharing the policer in input direction, one for all ports sharing the policer in output direction, one for all VLANs sharing the policer in input direction and one for all VLANs sharing the policer in output direction.

When creating a named aggregate policer, note the following:

• The valid range of values for the rate parameter is as follows:

– Minimum—32 kilobits per second

– Maximum—32 gigabits per second

For more information, see the “Configuration Guidelines” section.

• Rates can be entered in bits-per-second, or you can use the following abbreviations:

– k to denote 1000 bps

– m to denote 1000000 bps

– g to denote 1000000000 bps

Note You can also use a decimal point. For example, a rate of 1,100,000 bps can be entered
as 1.1 m.

• The valid range of values for the burst parameter is as follows:

– Minimum—1 kilobyte

– Maximum—512 megabytes

• Bursts can be entered in bytes, or you can use the following abbreviation:

– k to denote 1000 bytes

– m to denote 1000000 bytes

– g to denote 1000000000 bytes

Note You can also use a decimal point. For example, a burst of 1,100,000 bytes can be entered
as 1.1 m.

• Optionally, you can specify a conform action for matched in-profile traffic as follows:

– The default conform action is transmit.

– Enter the drop keyword to drop all matched traffic.

Note When you configure drop as the conform action, QoS configures drop as the exceed action.

• Optionally, for traffic that exceeds the CIR, you can specify an exceed action as follows:

– The default exceed action is drop.

– Enter the policed-dscp-transmit keyword to cause all matched out-of-profile traffic to be marked down as specified in the markdown map.

– For no policing, enter the transmit keyword to transmit all matched out-of-profile traffic.

• You can enter the no qos aggregate-policer policer_name command to delete a named aggregate policer.

This example shows how to create a named aggregate policer with a ten MBps rate limit and a one mb burst size that transmits conforming traffic and marks down out-of-profile traffic.

This example shows how to verify the configuration:

Configuring a QoS Policy

The following subsections describe QoS policy configuration:

• Overview of QoS Policy Configuration
• Configuring a Class Map (Optional)
• Configuring a Policy Map
• Attaching a Policy Map to an Interface

Note QoS policies process both unicast and multicast traffic.

Overview of QoS Policy Configuration

Configuring a QoS policy requires you to configure traffic classes and the policies that will be applied to those traffic classes, and to attach the policies to interfaces using these commands:

• access-list (optional for IP traffic—you can filter IP traffic with class-map commands):

– QoS supports these access list types:

– See “Configuring Network Security with ACLs,” for information about ACLs on the Catalyst 4500 series switches.

• class-map (optional)—Enter the class-map command to define one or more traffic classes by specifying the criteria by which traffic is classified. (For more information, see the “Configuring a Class Map (Optional)” section.)
• policy-map —Enter the policy-map command to define the following for each class of traffic:

– Internal DSCP source

– Aggregate or individual policing and marking

• service-policy —Enter the service-policy command to attach a policy map to an interface.

Configuring a Class Map (Optional)

The following subsections describe class map configuration:

• Creating a Class Map
• Configuring Filtering in a Class Map
• Verifying Class-Map Configuration

Enter the class-map configuration command to define a traffic class and the match criteria that will be used to identify traffic as belonging to that class. Match statements can include criteria such as an ACL, an IP precedence value, or a DSCP value. The match criteria are defined with one match statement entered within the class map configuration mode.

Creating a Class Map

To create a class map, enter this command:

Configuring Filtering in a Class Map

To configure filtering in a class map, enter one of these commands:

Note Any Input or Output policy that uses a class map with the match ip precedence or match ip dscp class-map commands, requires that you configure the port on which the packet is received to trust dscp. If not, the IP packet DSCP/IP-precedence is not used for matching the traffic; instead, the receiving port’s default DSCP is used. Starting with Cisco IOS Release 12.2(31)SG, the Supervisor Engine V-10GE allows you to classify traffic based on packet’s IP DSCP value irrespective of port trust state.

Note With Cisco IOS Release 12.2(31), the Catalyst 4500 series switch supports Match CoS.

Note The interfaces on the Catalyst 4500 family switch do not support the match classmap, match destination-address, match input-interface, match mpls, match not, match protocol, match qos-group, and match source-address keywords.

Verifying Class-Map Configuration

To verify class-map configuration, perform this task:

This example shows how to create a class map named ipp5 and how to configure filtering to match traffic with IP precedence 5:

This example shows how to verify the configuration:

This example shows how to configure match CoS for non-IPv4 traffic and how to configure filtering to match traffic with CoS value of 5:

This example shows how to verify the configuration:

Configuring a Policy Map

You can attach only one policy map to an interface. Policy maps can contain one or more policy-map classes, each with different match criteria and policers.

Configure a separate policy-map class in the policy map for each type of traffic that an interface receives. Put all commands for each type of traffic in the same policy-map class. QoS does not attempt to apply commands from more than one policy-map class to matched traffic.

The following sections describe policy-map configuration:

• Creating a Policy Map
• Configuring Policy-Map Class Actions

Creating a Policy Map

To create a policy map, enter this command:

Configuring Policy-Map Class Actions

These sections describe policy-map class action configuration:

• Configuring the Policy-Map Marking State
• Configuring the Policy-Map Class Trust State
• Configuring the Policy-Map Class DBL State
• Configuring Policy-Map Class Policing
• Using a Named Aggregate Policer
• Configuring a Per-Interface Policer

Configuring the Policy-Map Marking State

To configure the policy map to mark the IP precedence or dscp of a packet, enter this command:

Configuring the Policy-Map Class Trust State

To configure the policy-map class trust state, enter this command:

When configuring the policy-map class trust state, note the following:

• You can enter the no trust command to use the trust state configured on the ingress interface (the default).
• With the cos keyword, QoS sets the internal DSCP value from received or interface CoS.
• With the dscp keyword, QoS uses received DSCP.

Configuring the Policy-Map Class DBL State

To configure the policy-map class DBL state, enter this command:

When configuring the policy-map class DBL state, note the following:

• Any class that uses a named aggregate policer must have the same DBL configuration to work.

Configuring Policy-Map Class Policing

These sections describe configuration of policy-map class policing:

• Using a Named Aggregate Policer
• Configuring a Per-Interface Policer

Using a Named Aggregate Policer

To use a named aggregate policer (see the “Creating Named Aggregate Policers” section), enter this command:

Configuring a Per-Interface Policer

To configure a per-interface policer (see the “Policing and Marking” section), enter this command:

When configuring a per-interface policer, note the following:

• The valid range of values for the rate parameter is as follows:

– Minimum—32 kilobits per second, entered as 32000

– Maximum—32 gigabits per second, entered as 32000000000

Note For more information, see the “Configuration Guidelines” section.

• Rates can be entered in bits-per-second, or you can use the following abbreviations:

– k to denote 1000 bps

– m to denote 1000000 bps

– g to denote 1000000000 bps

Note You can also use a decimal point. For example, a rate of 1,100,000 bps can be entered
as 1.1m.

• The valid range of values for the burst parameter is as follows:

– Minimum—1 kilobyte

– Maximum—512 megabytes

• Bursts can be entered in bytes, or you can use the following abbreviation:

– k to denote 1000 bytes

– m to denote 1000000 bytes

– g to denote 1000000000 bytes

Note You can also use a decimal point. For example, a burst of 1,100,000 bytes can be entered
as 1.1m.

• Optionally, you can specify a conform action for matched in-profile traffic as follows:

– The default conform action is transmit.

– You can enter the drop keyword to drop all matched traffic.

• Optionally, for traffic that exceeds the CIR, you can enter the policed-dscp-transmit keyword to cause all matched out-of-profile traffic to be marked down as specified in the markdown map. See “Configuring the Policed-DSCP Map” section.

– For no policing, you can enter the transmit keyword to transmit all matched out-of-profile traffic.

This example shows how to create a policy map named ipp5-policy that uses the class map named ipp5. The class map ipp5 is configured to rewrite the packet precedence to 6 and to aggregate police the traffic that matches IP precedence value of 5:

This example shows how to create a policy map named cs2-policy that uses class map named cs2. The class map cos5 is configured to match on CoS 5 and to aggregate policing the traffic:

Verifying Policy-Map Configuration

To verify policy-map configuration, perform this task:

This example shows how to verify the configuration:

Attaching a Policy Map to an Interface

To attach a policy map to an interface, perform this task:

Note You cannot enable marking commands on an interface until IP routing is enabled globally. If IP routing is disabled globally and you try to configure the service policy on an interface, the configuration is accepted but it does not take effect. You are prompted with this message:

Set command will not take effect since CEF is disabled. Please enable IP routing and CEF globally.

To enable IP routing globally, enter the ip routing and ip cef global configuration commands. After you do this, the marking commands take effect.

This example shows how to attach the policy map named pmap1 to Fast Ethernet interface 5/36 and to verify the configuration:

Configuring CoS Mutation

Service providers providing Layer 2 VPNs carry double-tagged or Q-in-Q traffic with the outer tag representing the service providers’ VLAN, and the inner tag representing the customers’ VLAN. Differentiated levels of service can be provided inside the SP network based on the CoS present in the outer tag.

By using CoS mutation on a dot1q tunnel port, the CoS value on the outer tag of dot1q tunneled packets entering the provider core network can be derived from the CoS of the customer VLAN tag. This allows providers to preserve customer QoS semantics using their network.

CoS mutation is achieved through explicit user configuration to match on specific incoming CoS values and specifying the internal DSCP that is associated for matched packets. This internal DSCP gets converted to CoS through DSCP-CoS mapping during exit from the switch and is the CoS value that gets marked on the outer VLAN tag.

During the process, the CoS in inner tag is preserved and is carried across in the service providers’ network.

The following example shows how a policy map preserves customer VLAN IDs and CoS values throughout the network:

Configuring User-Based Rate-Limiting

User-Based Rate-Limiting (UBRL) adopts microflow policing capability to dynamically learn traffic flows and rate-limit each unique flow to an individual rate. UBRL is available on Supervisor Engine V-10GE with the built-in NetFlow support. UBRL can be applied to ingress traffic on routed interfaces with source or destination flow masks. It can support up to 85,000 individual flows and 511 rates. UBRL is typically used in environments where a per-user, granular rate-limiting mechanism is required; for example, the per-user outbound traffic rate could differ from the per-user inbound traffic rate.

Note By default, UBRL polices only routed IP traffic. You can use the ip flow ingress layer2-switched global command to police switched IP traffic. However, UBRL configuration must remain on a Layer 3 interface. With the UBRL configurations and the ip flow ingress layer2-switched global command, you will also be able to police intra-VLAN flows. You do not need to enter the ip flow ingress command.

A flow is defined as a five-tuple (IP source address, IP destination address, IP head protocol field, Layer 4 source, and destination ports). Flow-based policers enable you to police traffic on a per-flow basis. Because flows are dynamic, they require distinguishing values in the class map.

When you specify the match flow command with the source-address keyword, each flow with a unique source address is considered a new flow. When you specify the match flow command with the destination-address keyword, each flow with a unique destination address is considered a new flow. If the class map used by the policy map has any flow options configured, it is considered a flow-based policy map. When you specify the match flow command with the
ip destination-address ip protocol L4 source-address L4 destination-address keyword, each flow with unique IP source, destination, protocol, and Layer 4 source and destination address is considered a new flow.

Note Microflow is only supported on Supervisor Engine V-10GE.

To configure the flow-based class maps and policy maps, perform this task:

Examples

Example 1

This example shows how to create a flow-based class map associated with a source address:

Example 2

This example shows how to create a flow-based class map associated with a destination address:

Example 3

Assume there are two active flows on the Fast Ethernet interface 6/1 with source addresses 192.168.10.20 and 192.168.10.21. The following example shows how to maintain each flow to one MBps with an allowed burst value of 9000 bytes:

Example 4

Assume there are two active flows on the Fast Ethernet interface 6/1 with destination addresses of 192.168.20.20 and 192.168.20.21. The following example shows how to maintain each flow to one MBps with an allowed burst value of 9000 bytes:

Example 5

Assume that there are two active flows on FastEthernet interface 6/1:

With the following configuration, each flow is policed to 1000000 bps with an allowed 9000 burst value.

Note If you use the match flow ip source-address | destination-address command, these two flows are consolidated into one flow because they have the same source and destination address.

Configuring Hierarchical Policers

Note Hierarchical policers are only supported on Supervisor Engine V-10GE.

You can tie flow policers with the existing policers to create dual policing rates on an interface. For example, using dual policing, you can limit all incoming traffic rates on a given interface to 50 MBps and can limit the rate of each flow that is part of this traffic to two MBps.

You can configure hierarchical policers with the service-policy policy-map config command. A policy map is termed flow-based if the class map it uses matches any of the flow-based match criteria
(such as match flow ip source-address). Each child policy map inherits all the match access-group commands of the parent.

Note You can configure only flow-based policy maps as child policy maps. A parent policy map cannot be a flow-based policy map. Both the child policy map and parent policy map must have match-all in their class map configuration.

To configure a flow-based policy map as a child of an individual or aggregate policer, perform this task:

Note In a hierarchal policer configuration with parent as aggregate policer and child as microflow policer, child microflow policer matched packets report only the packets that are in the profile (that is, match the policing rate). Packets that exceed the policing rate are not reported in the class-map packet match statistics.

This example shows how to create a hierarchical policy map. A policy map with the name aggregate-policy has a class map with the name aggregate-class. A flow-based policy map with the name flow-policy is attached to this policy map as a child policy map.

In the following example, traffic in the IP address range of 101.237.0.0 to 101.237.255.255 is policed to 50 MBps. Flows ranging from 101.237.10.0 to 101.237.10.255 are individually policed to a rate of two MBps. This traffic goes through two policers: the aggregate policer and the other flow-based policer.

The following example shows the configuration for this scenario:

The following example shows how to verify the configuration:

Enabling Per-Port Per-VLAN QoS

The per-port per-VLAN QoS feature allows you to specify different QoS configurations on different VLANs on a given interface. Typically, you use this feature on trunk or voice VLANs (Cisco IP phone) ports, because they belong to multiple VLANs.

To configure per-port per-VLAN QoS, perform this task:

Example 1

Figure 37-5 displays a sample topology for configuring PVQoS. The trunk port gi3/1 is comprised of multiple VLANs (101 and 102). Within a port, you can create your own service policy per-VLAN. This policy, performed in hardware, might consist of ingress and egress policing, trusting DSCP, or giving precedence to voice packet over data.

Figure 37-5 Per-Port Per-VLAN Topology

The following configuration file shows how to perform ingress and egress policing per-VLAN using the policy map P31_QoS applied to port Gigabit Ethernet 3/1:

Example 2

Assume that interface Gigabit Ethernet 6/1 is a trunk port and belongs to VLANs 20, 300-301, and 400. The following example shows how to apply policy map p1 for traffic in VLANs 20 and 400 and policy map p2 to traffic in VLANs 300 through 301:

Example 3

The following command shows how to display policy map statistics on VLAN 20 configured on Gigabit Ethernet interface 6/1:

Example 4

The following command shows how to display policy map statistics on all VLANs configured on Gigabit Ethernet interface 6/1:

Enabling or Disabling QoS on an Interface

The qos interface command reenables any previously configured QoS features. The qos interface command does not affect the interface queueing configuration.

To enable or disable QoS features for traffic from an interface, perform this task:

This example shows how to disable QoS on interface VLAN 5:

This example shows how to verify the configuration:

Configuring VLAN-Based QoS on Layer 2 Interfaces

By default, QoS uses policy maps attached to physical interfaces. For Layer 2 interfaces, you can configure QoS to use policy maps attached to a VLAN. For more information, see the “Attaching a Policy Map to an Interface” section.

To configure VLAN-based QoS on a Layer 2 interface, perform this task:

Note If no input QoS policy is attached to a Layer 2 interface, then the input QoS policy attached to the VLAN (on which the packet is received), if any, is used even if the port is not configured as VLAN-based. If you do not want this default, attach a placeholder input QoS policy to the Layer 2 interface. Similarly, if no output QoS policy is attached to a Layer 2 interface, then the output QoS policy attached to the VLAN (on which the packet is transmitted), if any, is used even if the port is not configured as VLAN-based. If you do not want this default, attach a placeholder output QoS policy to the Layer 2 interface.

This example shows how to configure VLAN-based QoS on Fast Ethernet interface 5/42:

This example shows how to verify the configuration:

Note When a Layer 2 interface is configured with VLAN-based QoS, and if a packet is received on the port for a VLAN on which there is no QoS policy, then the QoS policy attached to the port, if any is used. This applies for both input and output QoS policies.

Configuring the Trust State of Interfaces

This command configures the trust state of interfaces. By default, all interfaces are untrusted.

To configure the trust state of an interface, perform this task:

When configuring the trust state of an interface, note the following:

• You can use the no qos trust command to set the interface state to untrusted.
• For traffic received on an ingress interface configured to trust CoS using the qos trust cos command, the transmit CoS is always the incoming packet CoS (or the ingress interface default CoS if the packet is received untagged).
• When the interface trust state is not configured to trust dscp using the qos trust dscp command, the security and QoS ACL classification always use the interface DSCP and not the incoming packet DSCP.
• Starting with Cisco IOS Release 12.2(31)SG, the Supervisor Engine V-10GE allows you to classify a packet based on the packet’s IP DSCP value irrespective of the port trust state. Packet transmit queuing is not impacted by this behavior. For information on transmit queues, refer to the “Configuring Transmit Queues” section.”

This example shows how to configure Gigabit Ethernet interface 1/1 with the trust cos keywords:

This example shows how to verify the configuration:

Configuring the CoS Value for an Interface

QoS assigns the CoS value specified with this command to untagged frames from ingress interfaces configured as trusted and to all frames from ingress interfaces configured as untrusted.

To configure the CoS value for an ingress interface, perform this task:

This example shows how to configure the CoS 5 as the default on Fast Ethernet interface 5/24:

This example shows how to verify the configuration:

Configuring DSCP Values for an Interface

QoS assigns the DSCP value specified with this command to non-IPv4 frames received on interfaces configured to trust DSCP and to all frames received on interfaces configured as untrusted.

To configure the DSCP value for an ingress interface, perform this task:

This example shows how to configure the DSCP 5 as the default on Fast Ethernet interface 5/24:

This example shows how to verify the configuration:

Configuring Transmit Queues

The following sections describe how to configure transmit queues:

• Mapping DSCP Values to Specific Transmit Queues
• Allocating Bandwidth Among Transmit Queues
• Configuring Traffic Shaping of Transmit Queues
• Configuring a High Priority Transmit Queue

Depending on the complexity of your network and your QoS solution, you might need to perform all of the procedures in the following sections. However, you will first need to answer the following questions:

• Which packets are assigned (by DSCP value) to each queue?
• What is the size of a transmit queue relative to other queues for a given port?
• How much of the available bandwidth is allotted to each queue?
• What is the maximum rate and burst of traffic that can be transmitted out of each transmit queue?

Independent of the QoS state of an interface, a switch ensures that all the transmit queues are enabled. Because the DSCP values are trusted by default, a switch uses the appropriate transmit queues based on the DSCP to map them. This queue selection is based on the internal DSCP to transmit queue mapping table.

Mapping DSCP Values to Specific Transmit Queues

To map the DSCP values to a transmit queue, perform this task:

This example shows how to map DSCP values to transit queue 2.

This example shows how to verify the configuration.

Allocating Bandwidth Among Transmit Queues

To configure the transmit queue bandwidth, perform this task:

The bandwidth rate varies with the interface.

For systems using Supervisor Engine II-Plus, Supervisor Engine II-Plus TS, Supervisor Engine III, and Supervisor Engine IV, bandwidth can be configured on these ports only:

• Uplink ports on supervisor engines
• Ports on the WS-X4306-GB GBIC module
• Ports on the WS-X4506-GB-T CSFP module
• The two 1000BASE-X ports on the WS-X4232-GB-RJ module
• The first two ports on the WS-X4418-GB module
• The two 1000BASE-X ports on the WS-X4412-2GB-TX module

For systems using Supervisor Engine V, bandwidth can be configured on all ports (10/100 Fast Ethernet, 10/100/1000BASE-T, and 1000BASE-X).

This example shows how to configure the bandwidth of one MBps on transmit queue 2:

Configuring Traffic Shaping of Transmit Queues

To guarantee that packets transmitted from a transmit queue do not exceed a specified maximum rate, perform this task:

This example shows how to configure the shape rate to one MBps on transmit queue 2:

Configuring a High Priority Transmit Queue

To configure transmit queue 3 at a higher priority, perform this task:

This example shows how to configure transmit queue 3 to high priority:

Configuring DSCP Maps

The following sections describes how to configure the DSCP maps. It contains this configuration information:

• Configuring the CoS-to-DSCP Map
• Configuring the Policed-DSCP Map
• Configuring the DSCP-to-CoS Map

All the maps are globally defined and are applied to all ports.

Configuring the CoS-to-DSCP Map

You use the CoS-to-DSCP map to map CoS values in incoming packets to a DSCP value that QoS uses internally to represent the priority of the traffic.

Table 37-3 shows the default CoS-to-DSCP map.

If these values are not appropriate for your network, you need to modify them.

To modify the CoS-to-DSCP map, perform this task:

This example shows how to configure the ingress CoS-to-DSCP mapping for CoS 0:

Note To return to the default map, use the no qos cos to dscp global configuration command.

This example shows how to clear the entire CoS-to-DSCP mapping table:

Configuring the Policed-DSCP Map

You use the policed-DSCP map to mark down a DSCP value to a new value as the result of a policing and marking action.

The default policed-DSCP map is a null map, which maps an incoming DSCP value to the same DSCP value.

To modify the CoS-to-DSCP map, perform this task:

To return to the default map, use the no qos dscp policed global configuration command.

This example shows how to map DSCP 50 to 57 to a marked-down DSCP value of 0:

Note In the previous policed-DSCP map, the marked-down DSCP values are shown in the body of the matrix. The d1 column specifies the most-significant digit of the original DSCP; the d2 row specifies the least-significant digit of the original DSCP. The intersection of the d1 and d2 values provides the marked-down value. For example, an original DSCP value of 53 corresponds to a marked-down DSCP value of 0.

Configuring the DSCP-to-CoS Map

You use the DSCP-to-CoS map to generate a CoS value.

Table 37-4 shows the default DSCP-to-CoS map.

If the values above are not appropriate for your network, you need to modify them.

To modify the DSCP-to-CoS map, perform this task:

This example shows how to map DSCP values 0, 8, 16, 24, 32, 40, 48, and 50 to CoS value 0 and to display the map:

Note In the previous DSCP-to-CoS map, the CoS values are shown in the body of the matrix. The d1 column specifies the most-significant digit of the DSCP; the d2 row specifies the least-significant digit of the DSCP. The intersection of the d1 and d2 values provides the CoS value. For example, in the DSCP-to-CoS map, a DSCP value of 08 corresponds to a CoS value of 0.

Generated Auto-QoS Configuration

By default, auto-QoS is disabled on all interfaces.

When you enable the auto-QoS feature on the first interface, these automatic actions occur:

• QoS is globally enabled (qos global configuration command).
• DBL is enabled globally (qos dbl global configuration command)
• When you enter the auto qos voip trust interface configuration command, the ingress classification on the specified interface is set to trust the CoS label received in the packet if the specified interface is configured as Layer 2 (and is set to trust DSCP if the interface is configured as Layer 3). See Table 37-5 .
• When you enter the auto qos voip cisco-phone interface configuration command, the trusted boundary feature is enabled. It uses the Cisco Discovery Protocol (CDP) to detect the presence or absence of a Cisco IP phone. When a Cisco IP phone is detected, the ingress classification on the interface is set to trust the CoS label received in the packet, if the interface is configured as Layer 2. (The classification is set to trust DSCP if the interface is configured as Layer 3.) When a Cisco IP phone is absent, the ingress classification is set to not trust the CoS label in the packet.

Note On a given port, the Cisco IP phone discovery information is not maintained on the standby supervisor engine. When the standby supervisor engine becomes active, it rediscovers the Cisco IP phone thru CDP. For a short time, the port are not in the trust state after the SSO switchover.

For information about the trusted boundary feature, see the “Configuring a Trusted Boundary to Ensure Port Security” section.

When you enable auto-QoS by using the auto qos voip cisco-phone or the auto qos voip trust interface configuration commands, the switch automatically generates a QoS configuration based on the traffic type and ingress packet label and applies the commands listed in Table 37-5 to the interface.

Effects of Auto-QoS on the Configuration

When auto-QoS is enabled, the auto qos voip interface configuration command and the generated configuration are added to the running configuration.

Configuration Guidelines

Before configuring auto-QoS, you should be aware of this information:

• In this release, auto-QoS configures the switch only for VoIP with Cisco IP phones.
• To take advantage of the auto-QoS defaults, do not configure any standard-QoS commands before entering the auto-QoS commands. If necessary, you can fine-tune the QoS configuration, but we recommend that you do so only after the auto-QoS configuration is completed.
• You can enable auto-QoS on static, dynamic-access, voice VLAN access, and trunk ports.
• By default, the CDP is enabled on all interfaces. For auto-QoS to function properly, do not disable the CDP.
• To enable auto qos voip trust on Layer 3 interfaces, change the port to Layer 3, then apply auto-QoS to make it trust DSCP.

Enabling Auto-QoS for VoIP

To enable auto-QoS for VoIP within a QoS domain, perform this task:

To disable auto-QoS on an interface, use the no auto qos voip interface configuration command. When you enter this command, the switch changes the auto-QoS settings to the standard-QoS default settings for that interface. It does not change any global configuration performed by auto-QoS. Global configuration remains the same.

This example shows how to enable auto-QoS and to trust the CoS labels in incoming packets when the device connected to Fast Ethernet interface 1/1 is detected as a Cisco IP phone:

This example shows how to enable auto-QoS and to trust the CoS/DSCP labels in incoming packets when the switch or router connected to Gigabit Ethernet interface 1/1 is a trusted device:

This example shows how to display the QoS commands that are automatically generated when auto-QoS is enabled:

Displaying Auto-QoS Information

To display the initial auto-QoS configuration, use the show auto qos [ interface interface-id ] privileged EXEC command. To d isplay any user changes to that configuration, use the show running-config privileged EXEC command. You can compare the show auto qos and the show running-config command output to identify the user-defined QoS settings.

To display information about the QoS configuration that might be affected by auto-QoS, use one of these commands:

• show qos
• show qos map
• show qos interface [ interface-id ]

For more information about these commands, refer to the command reference for this release.

Auto-QoS Configuration Example

This section describes how you could implement auto-QoS in a network, as shown in Figure 37-6.

Figure 37-6 Auto-QoS Configuration Example Network

The intelligent wiring closets in Figure 37-6 are composed of Catalyst 4500 switches. The object of this example is to prioritize the VoIP traffic over all other traffic. To do so, enable auto-QoS on the switches at the edge of the QoS domains in the wiring closets.

Note You should not configure any standard QoS commands before entering the auto-QoS commands. You can fine-tune the QoS configuration, but we recommend that you do so only after the auto-QoS configuration is completed.

To configure the switch at the edge of the QoS domain to prioritize the VoIP traffic over all other traffic, perform this task: