Catalyst 4500 Series Switch Software Configuration Guide, IOS XE 3.5.xE and IOS 15.2(1)Ex

Immediate-Leave Processing

IGMP snooping immediate-leave processing allows the switch to remove an interface from the forwarding-table entry without first sending out IGMP group-specific queries to the interface. The VLAN interface is pruned from the multicast tree for the multicast group specified in the original IGMP leave message. Immediate-leave processing ensures optimal bandwidth management for all hosts on a switched network, even when multiple multicast groups are being used simultaneously.

When a switch with IGMP snooping enabled receives an IGMPv2 or IGMPv3 leave message, it sends an IGMP group-specific query from the interface where the leave message was received to determine when other hosts are attached to that interface that are interested in joining the MAC multicast group. If the switch does not receive an IGMP join message within the query response interval, the interface is removed from the port list of the (MAC-group, VLAN) entry in the Layer 2 forwarding table.

Note By default all IGMP joins are forwarded to all multicast router ports.

With immediate-leave processing enabled on the VLAN, an interface can be removed immediately from the port list of the Layer 2 entry when the IGMP leave message is received, unless a multicast router was learned on the port.

Note When using IGMPv2 snooping, use immediate-leave processing only on VLANs where just one host is connected to each interface. If immediate-leave processing is enabled on VLANs where multiple hosts are connected to an interface, some hosts might be dropped inadvertently. When using IGMPv3, immediate-leave processing is enabled by default, and due to explicit host tracking, the switch can detect when a port has single or multiple hosts maintained by the switch for IGMPv3 hosts. As a result, the switch can perform immediate-leave processing when it detects a single host behind a given port.

Note IGMPv3 is interoperable with older versions of IGMP.

To display the IGMP version on a particular VLAN, use the show ip igmp snooping querier vlan command.

To display whether the switch supports IGMPv3 snooping, use the show ip igmp snooping vlan command.

To enable immediate-leave for IGMPv2, use the ip igmp snooping immediate-leave command.

Note Immediate-leave processing is enabled by default for IGMPv3.

IGMP Configurable-Leave Timer

Immediate-leave processing cannot be used on VLANs where multiple hosts may be connected to a single interface. To reduce leave latency in such a scenario, IGMPv3 provides a configurable leave timer.

In Cisco IOS Release 12.2(25)SG and earlier, the IGMP snooping leave time was based on query response time. If membership reports were not received by the switch before the query response time of the query expired, a port was removed from the multicast group membership.

In Cisco IOS Release 12.2(31)SG and later, you can configure the length of time that the switch waits after sending a group-specific query to determine if hosts are still interested in a specific multicast group. The IGMP leave response time can be configured from 100 to 5000 milliseconds. The timer can be set either globally or per-VLAN. The VLAN configuration of the leave time overrides the global configuration.

For configuration steps, see the “Configuring the IGMP Leave Timer” section.

IGMP Snooping Querier

IGMP Snooping Querier support was introduced in Cisco IOS Release 12.2(50)SG. This is a Layer 2 feature required to support IGMP snooping in a VLAN where PIM and IGMP are not configured because the multicast traffic does not require routing.

In a network where IP multicast routing is configured, the IP multicast router acts as the IGMP querier by sending general queries. If the IP-multicast traffic in a VLAN only needs to be Layer 2-switched, an IP-multicast router is not required. Without an IP-multicast router on the VLAN, you must configure another switch as the IGMP querier so that it can send queries.

When enabled, the IGMP snooping querier sends out periodic IGMPv2 queries that trigger IGMP report messages from the switch that requests IP multicast traffic. IGMP snooping listens to these IGMP reports to establish appropriate forwarding.

On switches that use IGMP to report interest in IP multicast traffic, configure at least one switch as the IGMP snooping querier in each supported VLAN.

You can configure a switch to generate IGMP queries on a VLAN regardless of whether IP multicast routing is enabled.

Explicit Host Tracking

Explicit host tracking (EHT) monitors group membership by tracking hosts that are sending IGMPv3 membership reports. This tracking enables a switch to detect host information associated with the groups of each port. EHT also enables the user to track the membership and various statistics.

EHT enables a switch to track membership on a per-port basis. Consequently, a switch is aware of the hosts residing on each port and can perform immediate-leave processing when there is only one host behind a port.

To determine whether EHT is enabled on a VLAN, use the show ip igmp snoop vlan command.

Default IGMP Snooping Configuration

Table 1-1 shows the IGMP snooping default configuration values.

Enabling IGMP Snooping Globally

To enable IGMP snooping globally, perform this task:

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

Enabling IGMP Snooping on a VLAN

To enable IGMP snooping on a VLAN, perform this task:

This example shows how to enable IGMP snooping on VLAN 2 and verify the configuration:

Configuring Learning Methods

The following sections describe IGMP snooping learning methods:

• Configuring PIM/DVMRP Learning
• Configuring CGMP Learning

Configuring PIM/DVMRP Learning

To configure IGMP snooping to learn from PIM/DVMRP packets, perform this task:

This example shows how to configure IP IGMP snooping to learn from PIM/DVMRP packets:

Configuring CGMP Learning

To configure IGMP snooping to learn from CGMP self-join packets, perform this task:

This example shows how to configure IP IGMP snooping to learn from CGMP self-join packets:

Configuring a Static Connection to a Multicast Router

To configure a static connection to a multicast router, enter the ip igmp snooping vlan mrouter interface command on the switch.

To configure a static connection to a multicast router, perform this task:

This example shows how to configure a static connection to a multicast router:

Enabling IGMP Immediate-Leave Processing

When you enable IGMP immediate-leave processing on a VLAN, a switch removes an interface from the multicast group when it detects an IGMPv2 leave message on that interface.

Note For IGMPv3, immediate-leave processing is enabled by default with EHT.

To enable immediate-leave processing on an IGMPv2 interface, perform this task:

This example shows how to enable IGMP immediate-leave processing on interface VLAN 200 and to verify the configuration:

Configuring the IGMP Leave Timer

Follows these guidelines when configuring the IGMP leave timer:

• You can configure the leave time globally or per-VLAN.
• Configuring the leave time on a VLAN overrides the global setting.
• The default leave time is 1000 milliseconds.
• The IGMP configurable leave time is only supported on hosts running IGMP Version 2.
• The actual leave latency in the network is usually the configured leave time. However, the leave time might vary around the configured time, depending on real-time CPU load conditions, network delays and the amount of traffic sent through the interface.

To enable the IGMP configurable-leave timer, perform this task:

This example shows how to enable the IGMP configurable-leave timer and to verify the configuration:

-----------------------------------

Configuring IGMP Snooping Querier

The IGMP Snooping Querier feature can be enabled either globally or per-VLAN.

Note The IGMP snooping querier is disabled by default.

To configure IGMP Snooping Querier, perform this task:

For an example of how to display Snooping Querier information, refer to the “Displaying IGMP Snooping Querier Information” section.

Configuring Explicit Host Tracking

For IGMPv3, EHT is enabled by default and can be disabled on a per-VLAN basis.

To disable EHT processing on a VLAN, perform this task:

This example shows how to disable IGMP EHT on VLAN 200 and to verify the configuration:

Configuring a Host Statically

Hosts normally join multicast groups dynamically, but you can also configure a host statically on an interface.

To configure a host statically on an interface, perform this task:

This example shows how to configure a host statically in VLAN 200 on interface Fast Ethernet 2/11:

Suppressing Multicast Flooding

An IGMP snooping-enabled switch floods multicast traffic to all ports in a VLAN when a spanning-tree topology change notification (TCN) is received. Multicast flooding suppression enables a switch to stop sending such traffic. To support flooding suppression, the following interface and global commands were introduced in Cisco IOS Release 12.1(11b)EW:

• [no | default] ip igmp snooping tcn flood (interface command)
• [no | default] ip igmp snooping tcn flood query count [1 - 10] (global command)
• [no | default] ip igmp snooping tcn query solicit (global command)

Prior to Cisco IOS Release 12.1(11b)EW, when a spanning tree topology change notification (TCN) was received by a switch, the multicast traffic was flooded to all the ports in a VLAN for a period of three IGMP query intervals. This was necessary for redundant configurations. In Cisco IOS Release 12.1(11b)EW, the default time period the switch waits before multicast flooding stops was changed to two IGMP query intervals.

This flooding behavior is undesirable if the switch that does the flooding has many ports that are subscribed to different groups. The traffic could exceed the capacity of the link between the switch and the end host, resulting in packet loss.

With the no ip igmp snooping tcn flood command, you can disable multicast flooding on a switch interface following a topology change. Only the multicast groups that have been joined by a port are sent to that port, even during a topology change.

With the ip igmp snooping tcn flood query count command, you can enable multicast flooding on a switch interface for a short period of time following a topology change by configuring an IGMP query threshold.

Typically, if a topology change occurs, the spanning tree root switch issues a global IGMP leave message (referred to as a “query solicitation”) with the group multicast address 0.0.0.0. When a switch receives this solicitation, it floods this solicitation on all ports in the VLAN where the spanning tree change occurred. When the upstream router receives this solicitation, it immediately issues an IGMP general query.

With the ip igmp snooping tcn query solicit command, you can now direct a non-spanning tree root switch to enter the same query solicitation.

The following sections provide additional details on the new commands and illustrate how you can use them.

IGMP Snooping Interface Configuration

A topology change in a VLAN may invalidate previously learned IGMP snooping information. A host that was on one port before the topology change may move to another port after the topology change. When the topology changes, the Catalyst 4500 series switch takes special actions to ensure that multicast traffic is delivered to all multicast receivers in that VLAN.

When the spanning tree protocol is running in a VLAN, a spanning tree topology change notification (TCN) is issued by the root switch in the VLAN. A Catalyst 4500 series switch that receives a TCN in a VLAN for which IGMP snooping has been enabled immediately enters into multicast flooding mode for a period of time until the topology restabilizes and the new locations of all multicast receivers are learned.

While in multicast flooding mode, IP multicast traffic is delivered to all ports in the VLAN, and not restricted to those ports on which multicast group members have been detected.

Starting with Cisco IOS Release 12.1(11b)EW, you can manually prevent IP multicast traffic from being flooded to a switch port by using the no ip igmp snooping tcn flood command on that port.

For trunk ports, the configuration applies to all VLANs.

By default, multicast flooding is enabled. Use the no keyword to disable flooding, and use default to restore the default behavior (flooding is enabled).

To disable multicast flooding on an interface, perform this task:

This example shows how to disable multicast flooding on interface Fast Ethernet 2/11:

IGMP Snooping Switch Configuration

By default, flooding mode persists until the switch receives two IGMP general queries. You can change this period of time by using the ip igmp snooping tcn flood query count n command, where n is a number between 1 and 10.

This command operates at the global configuration level.

The default number of queries is 2. The no and default keywords restore the default.

To establish an IGMP query threshold, perform this task:

This example shows how to modify the switch to stop flooding multicast traffic after four queries:

When a spanning tree root switch receives a topology change in an IGMP snooping-enabled VLAN, the switch issues a query solicitation that causes an Cisco IOS router to send out one or more general queries. The new command ip igmp snooping tcn query solicit causes the switch to send the query solicitation whenever it notices a topology change, even if that switch is not the spanning tree root.

This command operates at the global configuration level.

By default, query solicitation is disabled unless the switch is the spanning tree root. The default keyword restores the default behavior.

To direct a switch to send a query solicitation, perform this task:

This example shows how to configure the switch to send a query solicitation upon detecting a TCN:

Displaying Querier Information

To display querier information, perform this task:

This example shows how to display the IGMP snooping querier information for all VLANs on the switch:

This example shows how to display the IGMP snooping querier information for VLAN 3:

Displaying IGMP Host Membership Information

Note By default, EHT maintains a maximum of 1000 entries in the EHT database. Once this limit is reached, no additional entries are created. To create additional entries, clear the database with the
clear ip igmp snooping membership vlan command.

To display host membership information, perform this task:

This example shows how to display host membership information for VLAN 20 and to delete the EHT database:

This example shows how to display host membership for interface gi4/1:

This example shows how to display host membership for VLAN 20 and group 224.10.10.10:

Displaying Group Information

To display detailed IGMPv3 information associated with a group, perform one of the following tasks:

This example shows how to display the host types and ports of a group in VLAN 1:

This example shows how to display the current state of a group with respect to a source IP address:

This example shows how to display the current state of a group with respect to a host MAC address:

This example shows how to display summary information for an IGMPv3 group:

This example shows how to display the total number of group addresses learned by the system globally:

This example shows how to display the total number of group addresses learned on VLAN 5:

Displaying Multicast Router Interfaces

When you enable IGMP snooping, the switch automatically learns to which interface the multicast routers are connected.

To display multicast router interfaces, perform this task:

This example shows how to display the multicast router interfaces in VLAN 1:

Displaying MAC Address Multicast Entries

To display MAC address multicast entries for a VLAN, perform this task:

This example shows how to display MAC address multicast entries for VLAN 1:

This example shows how to display a total count of MAC address entries for VLAN 1:

Displaying IGMP Snooping Information on a VLAN Interface

To display IGMP snooping information on a VLAN, perform this task:

This example shows how to display IGMP snooping information on VLAN 5:

Displaying IGMP Snooping Querier Information

To display IGMP Snooping Querier information, perform this task:

This example shows how to display Snooping Querier information:

Using MVR in a Multicast Television Application

In a multicast television application, a PC or a television with a set-top box can receive the multicast stream. Multiple set-top boxes or PCs can be connected to one subscriber port, which is a switch port configured as an MVR receiver port. Figure 1-1 is an example configuration. DHCP assigns an IP address to the set-top box or the PC. When a subscriber selects a channel, the set-top box or PC sends an IGMP report to Switch A to join the appropriate multicast. If the IGMP report matches one of the configured IP multicast group addresses, the switch CPU modifies the hardware address table to include this receiver port and VLAN as a forwarding destination of the specified multicast stream when it is received from the multicast VLAN. Uplink ports that send and receive multicast data to and from the multicast VLAN are called MVR source ports.

Figure 1-1 Multicast VLAN Registration Example

When a subscriber changes channels or turns off the television, the set-top box sends an IGMP leave message for the multicast stream. The switch CPU sends a MAC-based general query through the receiver port VLAN. If there is another set-top box in the VLAN still subscribing to this group, that set-top box must respond within the maximum response time specified in the query. If the CPU does not receive a response, it eliminates the receiver port as a forwarding destination for this group.

Without Immediate Leave, when the switch receives an IGMP leave message from a subscriber on a receiver port, it sends out an IGMP query on that port and waits for IGMP group membership reports. If no reports are received in a configured time period, the receiver port is removed from multicast group membership. With Immediate Leave, an IGMP query is not sent from the receiver port on which the IGMP leave was received. Once the leave message is received, the receiver port is removed from multicast group membership, which speeds up leave latency. Enable Immediate Leave feature only on receiver ports to which a single receiver device is connected.

Because MVR multicast traffic is sent only on mVLANs, duplicating television-channel multicast traffic for subscribers on different VLANs is unnecessary. The IGMP leave and join messages are in the VLAN to which the subscriber port is assigned. The access layer switch (Switch A) modifies the forwarding behavior to allow traffic forwarding from the multicast VLAN to the subscriber port in a different VLAN. This is done by selectively allowing traffic to cross between the two VLANs.

IGMP reports are sent to the same IP multicast group address as the multicast data. The Switch A CPU must capture all IGMP join and leave messages from the receiver ports and forward them to the multicast VLAN of the source (uplink) port, based on the MVR mode.

Default MVR Configuration

Note Enabling the MVR command will set all the MVR default parameters.

Table 1-2 shows the default MVR configuration.

MVR Configuration Guidelines and Limitations

Follow these guidelines when configuring MVR:

• Ports can be configured as either a source port or a receiver port.

– Ports connected to subscribers are configured as receiver ports.

– Router ports or ports that are connected to another MVR switch are configured as source ports.

• Compatible mode

A source port configuration is required for those ports that must receive MVR traffic, even when there is no JOIN request from that port. All the MVR traffic received on the mVLAN on any port is forwarded to all source and receiver ports. (The receiver port should have been joined either by IGMP report or through static configurations).

• Dynamic mode

Source port configuration is not required, unless there is a port connected to another Layer 2 switch that runs MVR on the same mVLAN as this switch. Configure such ports as source ports. All MVR traffic received on the mVLAN on any ports is forwarded to the receiver or source ports that are joined either by IGMP report or through static configurations.

• Only one MVR VLAN can be configured.
• Although receiver ports that are connected to subscribers can be on different VLANs, they should not belong to the mVLAN.
• mRouter ports should not be configured as receiver ports.
• Both trunk and access ports can be configured as either source or receiver ports.
• The maximum number of MVR groups is fixed at 500.
• MVR cannot coexist with a PVLAN; do not configure MVR on a PVLAN.
• The IGMPSN group MAC address can alias with an MVR group's MAC address.

For example, 225.1.1.1 and 226.1.1.1 are IP addresses whose MAC addresses match to the same multicast MAC address (0100.5e01.0101). If 225.1.1.1 is configured as an MVR group then 225.1.1.1 is handled by MVR and 226.1.1.1 is handled by IGMPSN.

If the 226.1.1.1 host is present on the MVR trunk receiver, IGMPSN might not handle the forwarding for 226.1.1.1. Instead, the switch treats 226.1.1.1 as an MVR group and MVR handles forwarding on the mVLAN. You should not connect the hosts interested in MVR aliased groups on the MVR trunk receiver port. (By interested we mean that a host sends a JOIN request for a multicast group in order to receive the traffic or stream for that group.) This limitation applies only to MVR trunk receiver ports.

• MVR and multicast-routing cannot co-exist on the same switch. If you try to enable MVR while multicast routing or a multicast routing protocol are enabled, your operation is cancelled and you receive an error message. If you enable multicast routing or a multicast routing protocol while MVR is enabled, MVR is disabled and you receive a warning message.
• MVR can coexist with IGMP snooping on a switch.
• MVR is not supported with IPv6 multicast groups.
• MVR supports only IGMPv2 messages; MVR group reports derived from other versions are dropped.

Configuring MVR Global Parameters

If you select the default settings, you do not need to set the optional MVR parameters. If you want to change the default parameters (except for the MVR VLAN), you must first enable MVR.

To configure MVR parameters, perform these steps:

To return a switch to the default settings, use the no mvr [ mode | group ip-address | querytime | vlan ] global configuration commands.

The following example shows how to enable and verify MVR:

Note Fa2/1 is an mrouter port.

Configuring MVR on Access Ports

To configure the access port, perform these steps:

To return the interface to its default settings, use the no mvr [ type | immediate | vlan vlan-id | group ] interface configuration commands.

This example shows how to configure MVR "source and receiver" access ports:

Configuring MVR on a Trunk Port

To configure MVR on a trunk port, perform these steps:

This example shows how to configure MVR “receiver” VLANs on trunk ports:

Default IGMP Filtering Configuration

Table 1-4 shows the default IGMP filtering configuration.

Configuring IGMP Profiles

To configure an IGMP profile and to enter IGMP profile configuration mode, use the ip igmp profile global configuration command. From the IGMP profile configuration mode, you can specify the parameters of the IGMP profile to be used for filtering IGMP join requests from a port. When you are in IGMP profile configuration mode, you can apply these keywords:

• deny— Specifies that matching addresses are denied (the default condition).
• exit —Exits from igmp-profile configuration mode.
• no— Negates a command or sets its defaults.
• permit— Specifies that matching addresses are permitted.
• range— Specifies a range of IP addresses for the profile. You can enter a single IP address or a range with starting and ending addresses.

By default, no IGMP profiles are configured. When a profile is configured with neither the permit nor the deny keyword, the default is to deny access to the range of IP addresses.

To create an IGMP profile for a port, perform this task:

To delete a profile, use the no ip igmp profile profile-number global configuration command.

To delete an IP multicast address or range of IP multicast addresses, use the no range ip multicast address IGMP profile configuration command.

This example shows how to create IGMP profile 4 (allowing access to the single IP multicast address) and how to verify the configuration. If the action were to deny (the default), it does not appear in the output of the show ip igmp profile command.

Applying IGMP Profiles

To control access as defined in an IGMP profile, use the ip igmp filter interface configuration command to apply the profile to the appropriate interfaces. You can apply a profile to multiple interfaces, but each interface can only have one profile applied to it.

Note You can apply IGMP profiles to Layer 2 ports only. You cannot apply IGMP profiles to routed ports (or SVIs) or to ports that belong to an EtherChannel port group.

To apply an IGMP profile to a switch port, perform this task:

To remove a profile from an interface, use the no ip igmp filter command.

This example shows how to apply IGMP profile 4 to an interface and to verify the configuration:

Setting the Maximum Number of IGMP Groups

You can set the maximum number of IGMP groups that a Layer 2 interface can join by using the ip igmp max-groups interface configuration command. Use the no form of this command to set the maximum back to the default, which is no limit.

Note This restriction can be applied to Layer 2 ports only. You cannot set a maximum number of IGMP groups on routed ports (or SVIs) or on ports that belong to an EtherChannel port group.

To apply an IGMP profile on a switch port, perform this task:

This example shows how to limit the number of IGMP groups that an interface can join to 25: