In this post we will quickly discuss the use of most commonly needed IGMP timers. First, as we know, multicast routers periodically query hosts on a segment. If there are two or more routers sharing the same segment, the one with the lowest IP address is the IGMP querier (per IGMPv2 election procedure – as you remember, IGMPv1 let the multicast routing protocol define the querier).
The periodic interval is defined using the command:
ip igmp query-interval [interval in secs]
The hosts on the segment are supposed to report their group membership in response to the queries. Note that IGMPv2 defines special report suppression procedure, which allows host to suppress its own report, if it saw some other host reporting the same group. The query-interval timer is also used to define the amount of time a router will store particular IGMP state if it does not hear any reports on the group. This interval is 3xquery-time and it was the only mechanism available in IGMPv1 to expire a non-needed group. By default the interval is 60 seconds.
From what we said above follows that hosts must respond to a query during some time-window interval. This time window is specified in IGMPv2 packets using special field that defines the maximum response time. You set the value of this field using the command:
ip igmp query-max-response-time [time-in-seconds]
When a host receives the query packet, it starts counting to a random value, less that the maximum response time. When this timer expires, host replies with a report, provided that no other host has responded yet. This accomplishes two purposes:
a) Allows controlling the amount of IGMP reports sent during a time window.
b) Engages the report suppression feature, which permits a host to suppress its own report and conserve bandwidth.
Naturally, the maximum response timer could not be less than the query-interval. Note that IGMPv1 does not support the maximum response time field in its packets, and this timer is fixed to 10 seconds with version 1.
The next important timer, pertaining to IGMPv2, is last member query interval. This interval is configured using the following command:
ip igmp last-member-query-interval [interval in ms]
IGMP uses this value when router hears IGMPv2 Leave report. This means that at least one host wants to leave the group. After router receives the Leave report, it checks that the interface is not configured for IGMP Immediate Leave (single-host on the segment) and if not, it sends out an out-of-sequence query. The maximum-response-time in this query is set to last-member-query-interval which is 1000ms by default. The router sends out maximum of
ip igmp last-member-query-count [number]
messages and if no one responds during this time, the router removes the IGMP state for the group. The whole procedure controls if there are any more members left on the interface. After the last query send router waits some additional time, approximately 0,5 second to finally remove the group. So by default, after a reception of the Leave message the router waits for 2×1000ms+0,5seconds before stopping the multicast traffic flow.
If the interface is configured for immediate leave for a specific group list using the command:
ip igmp immediate-leave group-list [list]
Then the router will treat these groups as having single host member. After the reception of a Leave message, the router immediately removes the multicast forwarding state.
The last interesting timer is
ip igmp query-timeout [timeout]
This timer is used by “silent” routers, i.e. the routers that lost the IGMP querier election process. If the inactive routers did not hear any query for the 2 * [timeout] interval, they start election process again.
And in the end, a quick command to learn all default IGMP timer values:
By Petr Lapukhov, CCIE #16379
The periodic interval is defined using the command:
ip igmp query-interval [interval in secs]
The hosts on the segment are supposed to report their group membership in response to the queries. Note that IGMPv2 defines special report suppression procedure, which allows host to suppress its own report, if it saw some other host reporting the same group. The query-interval timer is also used to define the amount of time a router will store particular IGMP state if it does not hear any reports on the group. This interval is 3xquery-time and it was the only mechanism available in IGMPv1 to expire a non-needed group. By default the interval is 60 seconds.
From what we said above follows that hosts must respond to a query during some time-window interval. This time window is specified in IGMPv2 packets using special field that defines the maximum response time. You set the value of this field using the command:
ip igmp query-max-response-time [time-in-seconds]
When a host receives the query packet, it starts counting to a random value, less that the maximum response time. When this timer expires, host replies with a report, provided that no other host has responded yet. This accomplishes two purposes:
a) Allows controlling the amount of IGMP reports sent during a time window.
b) Engages the report suppression feature, which permits a host to suppress its own report and conserve bandwidth.
Naturally, the maximum response timer could not be less than the query-interval. Note that IGMPv1 does not support the maximum response time field in its packets, and this timer is fixed to 10 seconds with version 1.
The next important timer, pertaining to IGMPv2, is last member query interval. This interval is configured using the following command:
ip igmp last-member-query-interval [interval in ms]
IGMP uses this value when router hears IGMPv2 Leave report. This means that at least one host wants to leave the group. After router receives the Leave report, it checks that the interface is not configured for IGMP Immediate Leave (single-host on the segment) and if not, it sends out an out-of-sequence query. The maximum-response-time in this query is set to last-member-query-interval which is 1000ms by default. The router sends out maximum of
ip igmp last-member-query-count [number]
messages and if no one responds during this time, the router removes the IGMP state for the group. The whole procedure controls if there are any more members left on the interface. After the last query send router waits some additional time, approximately 0,5 second to finally remove the group. So by default, after a reception of the Leave message the router waits for 2×1000ms+0,5seconds before stopping the multicast traffic flow.
If the interface is configured for immediate leave for a specific group list using the command:
ip igmp immediate-leave group-list [list]
Then the router will treat these groups as having single host member. After the reception of a Leave message, the router immediately removes the multicast forwarding state.
The last interesting timer is
ip igmp query-timeout [timeout]
This timer is used by “silent” routers, i.e. the routers that lost the IGMP querier election process. If the inactive routers did not hear any query for the 2 * [timeout] interval, they start election process again.
And in the end, a quick command to learn all default IGMP timer values:
Rack21R5#show ip igmp interface fastEthernet 0/1 FastEthernet0/1 is up, line protocol is up Internet address is 129.21.59.5/24 IGMP is enabled on interface Current IGMP host version is 2 Current IGMP router version is 2 IGMP query interval is 60 seconds IGMP querier timeout is 120 seconds IGMP max query response time is 10 seconds Last member query count is 2 Last member query response interval is 1000 ms Inbound IGMP access group is not set
By Petr Lapukhov, CCIE #16379
No comments:
Post a Comment