Hosts participating in IP multicast can join and leave a multicast group at any time. The number and locations of members in a multicast group are restricted. A multicast router unnecessarily and also impossibly keeps the membership of all hosts. It only uses IGMP to learn whether receivers (namely, group members) in a multicast group are present on the subnet connected to each interface. A host only needs to keep the multicast groups which it has joined.

IGMP is not symmetric on hosts and routers. Hosts need to respond to IGMP query messages from the multicast router, that is, report the group membership to the router. The router needs to send membership query messages periodically to discover whether hosts join the specified group on its subnets according to the received response messages. When the router receives the report that hosts leave the group, the router will send a group-specific query packet (IGMP Version 2) to discover whether any member exists in the group.

Up to now, IGMP has three versions, namely, IGMP Version 1 (defined by RFC1112), IGMP Version 2 (defined by RFC2236) and IGMP Version 3. IGMP Version 2 is the most widely used version.

IGMP Version 2 boasts the following improvements over IGMP Version 1:

I. Election mechanism of multicast routers on the shared network segment

A shared network segment means that there are multiple multicast routers on a network segment. In this case, all routers running IGMP on the network segment can receive the membership report from hosts. Therefore, only one router is necessary to send membership query messages. In this case, the router election mechanism is required to specify a router as the querier.

In IGMP Version 1, selection of the querier is determined by the multicast routing protocol. While IGMP Version 2 specifies that the multicast router with the lowest IP address is elected as the querier when there are multiple multicast routers on the same network segment.

II. Leaving group mechanism

In IGMP Version 1, hosts leave the multicast group quietly without informing the multicast router. In this case, the multicast router can determine whether hosts leave the group according to the timeout of responses to the multicast group. In Version 2, when a host is intended to leave, it will send a leave group message if it is the host which responds to the latest membership query message.

III. Specific group query

In IGMP Version 1, a query of a multicast router is targeted at all the multicast groups on the network segment, which is known as General Query.

In IGMP Version 2, Group-Specific Query is added besides general query. The destination IP address of the query packet is the IP address of the multicast group. The group address domain in the packet is also the IP address of the multicast group. This prevents hosts in other multicast groups from sending response messages.

IV. Max response time

The Max Response Time is added in IGMP Version 2. It is used to dynamically adjust the allowed maximum time for a host to respond to the group query message.

1.2 IGMP Proxy

1.2.1 Introduction to IGMP Proxy

For a large scale network with multicast routing protocol employed, many leaf networks may exist (a leaf network here refers to an end node of a multicast forwarding tree, it is a subnet to which only multicast clients are attached). It is a heavy load to configure and manage all these leaf networks.

You can ease the workload of configuring and managing leaf networks without affecting the multicast connections in them by enabling IGMP proxy on devices in these leaf networks.

After IGMP proxy is configured, the devices in leaf networks act as a host to the exterior network. They receive multicast data of the associated group only when some of the hosts directly connected to them are multicast group members.

I. Configuration description

Figure 1-1 A schematic diagram of IGMP proxy

Figure 1-1 illustrates how IGMP proxy works.

Switch B is configured as follows:

l Multicast is enabled.

l PIM and IGMP are configured on VLAN-interface 100 and VLAN-interface 200.

l VLAN-interface 100 is configured as the IGMP proxy interface of VLAN-interface 200.

Switch A is configured as follows:

l Multicast is enabled.

l PIM and IGMP are configured on VLAN-interface 100.

l The pim neighbor-policy command is executed on VLAN-interface100 to filter PIM neighbors on the network segment 33.33.33.0/24. That is, prevent Switch B from being a PIM neighbor.

II. Operating mechanism

The procedure for processing IGMP join/leave messages is as follows:

l After receiving an IGMP join/leave message sourced from a host through VLAN–interface 200, Switch B changes the source address of the message to the IP address of VLAN–interface 100 (33.33.33.2), which is the outbound interface to Switch A.

l Switch B sends the IGMP message to Switch A.

l Switch A processes the message after receiving the IGMP message sent by Switch B through VLAN-interface 100, just as the message is sent by a host directly connected to VLAN-interface 100.

The procedures to process IGMP normal group or specific group querying messages are as follows:

l After receiving a normal group or a specific group querying message from Switch A, Switch B changes the source address of the querying message to the address of the outbound interface leading to hosts.

l Switch B transmits the message through VLAN-interface 200.

1.2.2 IGMP Proxy Configuration

I. Configuration prerequisites

l IP addresses, PIM, and PIM neighbor-policy have been configured on the corresponding interfaces of the devices in the exterior network.

l The switches in the leaf networks are multicast-enabled. The related ports are added to specific VLANs. IP addresses, PIM, and IGMP have been configured on the corresponding interfaces.

II. Configuration procedure

Follow these steps to configure IGMP proxy:

To do...	Use the command...	Remarks
Enter system view	system-view	—
Enter interface view	interface interface-type interface-number	—
Enable IGMP Proxy on the current interface and specify the proxy interface for the current interface	igmp proxy interface-type interface-number	Required By default, IGMP Proxy is disabled on the interface

Caution:

l Enable the IGMP protocol on the interface before configuring IGMP Proxy.

l If IGMP is not enabled on the specified interface, you can configure IGMP, but it does not take effect.

l If you configure IGMP Proxy on an interface for multiple times, the latest configuration is valid.

l One interface cannot be the IGMP proxy interface of two or more other interfaces simultaneously.

1.3 IGMP Configuration

Caution:

After multicast routing is enabled, you must enable IGMP on the interface before you can perform other IGMP configurations.

1.3.1 IGMP Basic Configuration

IGMP basic configuration includes:

l Enabling multicast routing

l Enabling IGMP on an interface

I. Enabling multicast routing

Multicast routing must be enabled before Layer 3 multicast-related configurations, such as IGMP and PIM configurations, can take effect.

For details, refer to Common Multicast Configuration in the IP Multicast Volume.

II. Enabling IGMP on an interface

This configuration task is to enable IGMP on the interface which needs to maintain the multicast membership. After this, you can perform IGMP feature configuration.

Perform the following operations in interface view to enable/disable IGMP on an interface:

To do...	Use the command...
Enable IGMP on the current interface	igmp enable
Disable IGMP on the current interface	undo igmp enable

By default, IGMP is disabled on the interface.

Caution:

l If the VLAN VPN is enabled on a port, the IGMP Snooping feature cannot be enabled on the VLAN to which the port belongs, and the IGMP feature cannot be enabled on the corresponding interface either.

l If IGMP Snooping feature is enabled on a VLAN, or IGMP is enabled on the interface, you cannot add a VPN-enabled port into the VLAN.

1.3.2 IGMP Advanced Configuration

IGMP advanced configuration includes:

l Enabling compatibility control function of the switch

l Configuring the IGMP version

l Configuring IGMP query interval

l Configuring the interval and the times of sending IGMP group-specific queries

l Configuring the present time of IGMP querier

l Configuring maximum response time for IGMP query message

l Limiting the number of multicast groups on an interface

l Configuring a router to join a multicast group

l Deleting IGMP groups joined on an interface

l Configuring the filtering rule of multicast groups

l Enabling/disabling IGMP fast leave processing

I. Enabling compatibility control function of the switch

With the compatibility control function enabled the switch processes the protocol message with the destination IP address 224.0.0.1 in the IGMP Report packet. Otherwise, the switch discards this kind of packets.

Perform the following operation in system view to enable the compatibility control function of the switch:

To do...	Use the command...
Enable compatibility control function of the switch	igmp-report enhance enable

By default, the compatibility control function of the switch is disabled.

This command is often executed after IGMP or IGMP Spooning protocol is enabled.

II. Configuring the IGMP version

Perform the following operations in VLAN interface view to configure the IGMP version:

To do...	Use the command...
Select the IGMP version	igmp version { 1 \| 2 }
Restore the default setting	undo igmp version

The default IGMP version is version 2 on S9500 series switches.

Caution:

The system does not support automatic switching between different IGMP versions. Therefore, all devices on a subnet must be configured to run the same IGMP version.

III. Configuring IGMP query interval

Multicast routers send IGMP query messages to discover which multicast groups have members on the local subnet. A multicast router sends query messages periodically to refresh their knowledge of multicast members present on the local subnet.

Perform the following configuration to configure the interval to send IGMP query message in VLAN interface view:

To do...	Use the command...
Configure the interval between IGMP query messages	igmp timer query seconds
Restore the default value	undo igmp timer query

When there are multiple multicast routers on a network segment, the querier is responsible for sending IGMP query messages to all hosts on the LAN.

By default, the interval is 60 seconds.

Caution:

The newly configured IGMP query interval has effect on those multicast groups that receive new reports after this configuration. Other multicast groups do not conform to the new aging time until they receive new reports.

IV. Configuring the interval and the times of sending IGMP group-specific queries

On the shared network, it is the query router (querier) that maintains IGMP membership on the interface. The igmp lastmember-queryinterval and igmp robust-count commands are used to configure the interval and times of sending multicast group-specific query packets for the querier when it receives an IGMP leave message from a host.

l The host sends the IGMP Leave message.

l Upon receiving the message, IGMP querier sends the group-specific IGMP query message for specified times (defined by the robust-value in igmp robust-count, with the default value being 2) and at a time interval (defined by the seconds in igmp lastmember-queryinterval, with the default value being 1 second).

l When other hosts receive the message from the IGMP querier and are interested in this group, they return the IGMP Membership Report message within the defined maximum response time.

l If IGMP querier receives the report messages from other hosts within the period equal to robust-value ´ seconds, it continues membership maintenance for this group.

l If it receives no report message from any other host within this period, it reckons this as timeout and ends membership maintenance for this group.

This configuration takes effect only when the querier runs IGMP version 2. If a host runs IGMP Version 1, it does not send IGMP Leave Group message when it leaves a group. In this case, this configuration does not work for the host.

1) Configuring interval for querying IGMP packets

Perform the following configuration to configure interval for sending IGMP group-specific queries in VLAN interface view:

To do...	Use the command...
Configure interval for sending IGMP group-specific queries	igmp lastmember-queryinterval seconds
Restore the default setting	undo igmp lastmember-queryinterval

By default, the interval is 1 second.

2) Configuring the times of sending IGMP group-specific queries

Perform the following configuration to configure the times of sending IGMP group-specific queries in VLAN interface view:

To do...	Use the command...
Configure number of last member querying	igmp robust-count robust-value
Restore the default number of querying	undo igmp robust-count

By default, a multicast group-specific query message is sent for twice.

V. Configuring the present time of IGMP querier

On a shared network, namely a network segment where multiple multicast routers exist, a query router (querier for short) sends query messages on the interface regularly. If a non-query router fails to receive messages from the querier within a period of time, it will deem that the querier has failed and take over the job of the original querier.

In the IGMP V1 version, the querier selection is determined by the multicast routing protocol; in the IGMP V2 version, the router with the smallest IP address on a shared network segment acts as the querier.

The IGMP querier presence time is the period of time before the router takes over as the querier sending query messages, after the previous querier has stopped doing so.

Perform the following configuration to configure the present time of IGMP querier in VLAN interface view:

To do...	Use the command...
Change the present time of IGMP querier	igmp timer other-querier-present seconds
Restore the default value	undo igmp timer other-querier-present

By default, the value is twice the IGMP query message interval, namely 120 seconds.

VI. Configuring maximum response time for IGMP query message

When a router receives a query message, the host will set a timer for each multicast group it belongs to. The value of the timer is randomly selected between 0 and the maximum response time. When any timer becomes 0, the host will send the membership report message of the multicast group.

Setting the maximum response time reasonably can enable the host to respond to query messages quickly. In this case, the router can fast master the existing status of the members of the multicast group.

Perform the following configuration to configure the maximum response time for IGMP query message in interface view:

To do...	Use the command...
Configure the maximum response time for IGMP query message	igmp max-response-time seconds
Restore the maximum query response time to the default value	undo igmp max-response-time

The smaller the maximum query response time value, the faster the router prunes groups. The actual response time is a random value in the range from 1 to 25 seconds. By default, the maximum query response time is 10 seconds.

VII. Limiting the number of multicast groups on an interface

If there is no limit to the number of multicast groups added on a router interface or a router, the router memory may be exhausted, which may cause router failure.

You can set the maximum number of multicast groups added on the interface, but not the maximum number of multicast groups allowed on the router, which is defined by the system.

Perform the following configuration to configure the maximum number of multicast groups on an interface in VLAN interface view:

To do...	Use the command...
Configure the maximum number of multicast groups on the interface	igmp group-limit limit
Restore the maximum number of multicast groups on the interface to the default value	undo igmp group-limit

By default, the maximum number of multicast groups on an interface is 512.

& Note:

This command can be successfully executed only when the number of multicast groups on the current interface is less than the specified upper limit.

VIII. Configuring a router to join a multicast group

Usually, the host operating IGMP will respond to IGMP queries of the multicast router. In case of response failure, the multicast router will consider that there is no multicast member on this network segment and will cancel the corresponding path. Configuring one interface of the router as multicast member can avoid such problem. When the interface receives IGMP query packet, the router will respond, thus ensuring that the network segment where the interface located can normally receive multicast packets.

For an Ethernet switch, you can configure a port in a switch interface to join a multicast group.

Perform the following configuration to configure a router to join a multicast group in the corresponding view:

To do...	Use the command...
Configure the router to join a specified multicast group (in VLAN interface view)	igmp host-join group-address port interface-type interface-number [ to interface-type interface- number ]
Cancel the configuration (in VLAN interface view)	undo igmp host-join group-address port interface-type interface-num [ to interface-type interface-number ]
Configure the router to join a specified multicast group (in Ethernet port view)	igmp host-join group-address vlan vlan-id
Cancel the configuration (in Ethernet port view)	undo igmp host-join group-address vlan vlan-id

& Note:

The above two configuration methods have the same result. You can select either of them.

By default, a router joins no multicast group. Note that the specified port must belong to this interface, and IGMP is enabled on this interface or IGMP Snooping is enabled in the VLAN. Otherwise, the configuration does not take effect.

Caution:

Secondary VLANs do not support static multicast groups, but you can configure static multicast groups in an isolate-user-VLAN, which has the same effect as configured in a secondary VLAN.

IX. Deleting IGMP groups joined on an interface

This configuration task is to delete all multicast groups joined on all interfaces or specific interfaces of the router, or to delete the multicast groups at the specific address or in the specific network segment on the specific interfaces of the router.

Perform the following configuration to delete multicast groups joined on an interface in user view:

To do...	Use the command...
Delete multicast groups joined on an interface	reset igmp group { all \| interface vlan-interface interface-number { all \| group-address [ group-mask ] } }

After a group is deleted, if other IGMP membership report messages occur, the interfaces can join the corresponding group again.

X. Configuring the filtering rule of multicast groups

On the IGMP snooping-enabled switch, you can configure ACL rules whether the specified multicast group can be joined to a VLAN or not. This feature filters every received IGMP join packet. According to the destination group address of the packets and the ACL rule bound to the VLAN, the switch determines whether to discard the packets or let them pass.

By setting the filtering rule of multicast groups in the VLAN, you can control access to multicast groups. You can only configure one ACL rule for each VLAN, and the new configured rule will replace the old one.

Perform the following configuration to configure the aging time of multicast group members in VLAN view:

To do...	Use the command...
Set the filtering rule of multicast groups in the specified VLAN	igmp-snooping group-policy acl-number
Cancel the filtering rule of multicast groups in the specified VLAN	undo igmp-snooping group-policy

By default, no filtering rule is set for a VLAN. In this case, a host can be joined to any multicast group.

After the ACL for filtering multicast groups is configured on a VLAN, the multicast groups that are not permitted by the ACL cannot be created.

Caution:

l If an inexistent ACL is bound to the VLAN, or if the bound ACL is not configured with a rule, a host is not allowed to join any multicast group.

l The multicast group filtering rule is not restricted by the ACL itself, and is valid for all members in the specified VLAN.

l The rules for filtering multicast groups have no effect on static multicast groups configured locally.

XI. Enabling/disabling IGMP fast leave processing

On an IGMP-enabled Layer 3 switch, when the fast leave-enabled port receives an IGMP leave packet, the switch removes the port from the outbound port lists of all Layer 3 multicast forwarding tables that are of the same multicast group to peel off the port from the multicast group. That is, the switch does not forward multicast data to the port.

Perform the following configuration to enable/disable IGMP fast leave processing in Ethernet port view or system view:

To do...	Use the command...	Remarks
Enter system view	system-view	—
Enable IGMP fast leave in system view	igmp fast-leave [ vlan { vlan-id [ to vlan-id ] } &<1-10> ]	IGMP fast leave is disabled by default
Enter Ethernet port view	interface interface-type interface-number	—
Enable IGMP fast leave in Ethernet port view	igmp fast-leave [ vlan { vlan-id [ to vlan-id ] } &<1-10> ]	IGMP fast leave is disabled by default
Disable IGMP fast leave	undo igmp fast-leave [ vlan { vlan-id [ to vlan-id ] } &<1-10> ]	—

For detailed configuration, refer to the corresponding commands.

& Note:

The configuration made in system view will be effective to ports within all the specified VLANs, while the configuration in port view will be effective to the port within the specific VLANs (for example, when a trunk port belongs to multiple VLANs).

Caution:

l If the specified VLANs do not exists, the port does not belongs to any of the specified VLANs, or the VLANs do not have IGMP enabled, you can still configure the fast leave feature, but the configuration will not take effect.

l You must enable multicast routing globally by executing the multicast routing-enable command in system view before you can configure the fast leave feature.

l If global multicast routing is disabled, all existing IGMP fast leave-related configurations will be cleared.

l When you configure IGMP fast leave on aggregation ports, the configuration takes effect only on primary aggregation ports.

l If you have added an IGMP V1 host of the same multicast group to the port, or configured a static host of the same multicast group by using the igmp host-join command, the switch does not remove the port when the port receives an IGMP Leave packet of the multicast group even if you have enabled IGMP fast leave for the port.

1.4 Displaying and Maintaining IGMP

To do…	Use the command…	Remarks
View IGMP multicast group information	display igmp group [ group-address \| interface vlan-interface interface-number ]	Available in any view
View IGMP configuration and operation information on an interface	display igmp interface [ vlan-interface interface-number ]	Available in any view
Enable IGMP debugging	debugging igmp { all \| event \| host \| packet \| timer }	Available in any view
Disable IGMP debugging	undo debugging igmp { all \| event \| host \| packet \| timer }	Available in any view

1.5 IGMP Configuration Examples

I. Network requirements

As shown in Figure 1-2, Switch B resides in a leaf network. Configure IGMP proxy for Switch B to ease the configuration and management work load in the leaf network without affecting multicast connections in it.

You need to perform the following configurations to meet the requirements:

l Enable IGMP and PIM-DM for the related VLAN interfaces on Switch A.

l Enable multicast on Switch B. Enable PIM for the interfaces of VLAN 100 and VLAN 200. Configure VLAN-interface 100 as the proxy interface of VLAN-interface 200.

& Note:

The following depicts IGMP and IGMP proxy configuration (other related configuration is not covered here).

II. Network diagram

Figure 1-2 Network diagram for IGMP proxy configuration

III. Configuration steps

1) Configure Switch B.

# Enable multicast.

<SwitchB> system-view

System View: return to User View with Ctrl+Z.

[SwitchB] multicast routing-enable

# Create VLAN 200 and VLAN 100. Add related ports to the corresponding VLANs. For details, refer to VLAN Configuration in the Access Volume.

# Enable IGMP and PIM-DM on VLAN-interface 200 and VLAN-interface 100.

[SwitchB] interface vlan-interface 200

[SwitchB-Vlan-interface 200] ip address 22.22.22.1 24

[SwitchB-Vlan-interface 200] igmp enable

[SwitchB-Vlan-interface 200] pim dm

[SwitchB-Vlan-interface 200] igmp host-join 224.0.1.1 port Ethernet 3/1/8

[SwitchB] interface vlan-interface 100

[SwitchB-Vlan-interface 100] ip address 33.33.33.2 24

[SwitchB-Vlan-interface 100] igmp enable

[SwitchB-Vlan-interface 100] pim dm

[SwitchB-Vlan-interface 100] quit

# Configure VLAN-interface 100 to be the IGMP proxy interface of VLAN-interface 200.

[SwitchB] interface vlan-interface 200

[SwitchB-Vlan-interface 200] igmp proxy Vlan-interface 100

2) Configure Switch A.

# Enable multicast.

<SwitchA> system-view

System View: return to User View with Ctrl+Z.

[SwitchA] multicast routing-enable

# Enable IGMP and PIM-DM for VLAN-interface 100.

[SwitchA] interface vlan-interface 100

[SwitchA-Vlan-interface 100] igmp enable

[SwitchA-Vlan-interface 100] pim dm

# Configure Vlan-interface 100 so that it will not use the IP address 33.33.33.2 as a PIM neighbor

[SwitchA-Vlan-interface 100] pim neighbor-policy 2001

[SwitchA-Vlan-interface 100] quit

[SwitchA] acl number 2001

[SwitchA-acl-basic-2001] rule deny source 33.33.33.2 0

[SwitchA-acl-basic-2001] rule permit source any

3) Configure Receiver

Receiver establishes HTTP connections to the multicast source and sends the list of the names of the services provided by the multicast source to the clients (Each service has its own multicast address, which is the multicast group). When a service is selected, the client sends IGMP packets to join the multicast group, through which the receiver can receive the corresponding multicast packets.

H3C S9500 Operation Manual-Release1648[v1.24]-04 IP Multicast Volume

1.1 IGMP Overview

I. Election mechanism of multicast routers on the shared network segment

II. Leaving group mechanism

III. Specific group query

IV. Max response time

I. Configuration description

II. Operating mechanism

I. Configuration prerequisites

II. Configuration procedure

II. Enabling IGMP on an interface

I. Enabling compatibility control function of the switch

II. Configuring the IGMP version

IV. Configuring the interval and the times of sending IGMP group-specific queries

V. Configuring the present time of IGMP querier

VI. Configuring maximum response time for IGMP query message

VII. Limiting the number of multicast groups on an interface

VIII. Configuring a router to join a multicast group

IX. Deleting IGMP groups joined on an interface

X. Configuring the filtering rule of multicast groups

XI. Enabling/disabling IGMP fast leave processing

1.5 IGMP Configuration Examples

I. Network requirements

II. Network diagram

III. Configuration steps

Intelligent Terminal Products

Product Support Services

Technical Service Solutions

Resource Center

Policy

Online Help

Become a Partner

Partner Policy & Program

Global Learning

Partner Sales Resources

Service Business

Company Information

News & Events

Contact Us