Triple Play multicast

Internet Group Management Protocol (IGMP) is used by IPv4 hosts and routers to report their IP multicast group memberships to neighboring multicast routers. A multicast router keeps a list of multicast group memberships for each attached network, and a timer for each membership.

Multicast group memberships include at least one member of a multicast group on a given attached network, not a list of all the members. With respect to each of its attached networks, a multicast router can assume one of two roles, querier or non-querier. There is normally only one querier per physical network.

A querier issues two types of queries, a general query and a group-specific query. General queries are issued to solicit membership information about any multicast group. Group-specific queries are issued when a router receives a leave message from the node it perceives as the last group member remaining on that network segment.

Hosts wanting to receive a multicast session issue a multicast group membership report. These reports must be sent to all multicast enabled routers.

Multicast Listener Discovery (MLD) is used by IPv6 hosts and routers to report their IP multicast group memberships to neighboring multicast routers. A multicast router keeps a list of multicast group memberships for each attached network, and a timer for each membership. Multicast group memberships include at least one member of a multicast group on a specific attached network, not a list of all the members. With respect to each of its attached networks, a multicast router can assume one of two roles, querier or non-querier. There is normally only one querier per physical network.

A querier issues two types of queries, a general query and a group-specific query. General queries are issued to solicit membership information about any multicast group. Group-specific queries are issued when a router receives a leave message from the node it perceives as the last group member remaining on that network segment.

Hosts wanting to receive a multicast session issue a multicast group membership report. These reports must be sent to all multicast enabled routers.

Protocol Independent Multicast Sparse Mode (PIM-SM) leverages the unicast routing protocols that are used to create the unicast routing table: OSPF, IS-IS, BGP, and static routes. Because PIM uses this unicast routing information to perform the multicast forwarding function it is effectively IP protocol independent. Unlike DVMRP, PIM does not send multicast routing tables updates to its neighbors.

PIM-SM uses the unicast routing table to perform the Reverse Path Forwarding (RPF) check function instead of building up a completely independent multicast routing table.

PIM-SM only forwards data to network segments with active receivers that have explicitly requested the multicast group. PIM-SM in the ASM model initially uses a shared tree to distribute information about active sources. Depending on the configuration options, the traffic can remain on the shared tree or switch over to an optimized source distribution tree. As multicast traffic starts to flow down the shared tree, routers along the path determine if there is a better path to the source. If a more direct path exists, then the router closest to the receiver sends a join message toward the source and then reroutes the traffic along this path.

As stated above, PIM-SM relies on an underlying topology-gathering protocol to populate a routing table with routes. This routing table is called the Multicast Routing Information Base (MRIB). The routes in this table can be taken directly from the unicast routing table, or it can be different and provided by a separate routing protocol such as MBGP. Regardless of how it is created, the primary role of the MRIB in the PIM-SM protocol is to provide the next hop router along a multicast-capable path to each destination subnet. The MRIB is used to determine the next hop neighbor to whom any PIM join/prune message is sent. Data flows along the reverse path of the join messages. Thus, in contrast to the unicast RIB that specifies the next hop that a data packet would take to get to some subnet, the MRIB gives reverse-path information, and indicates the path that a multicast data packet would take from its origin subnet to the router that has the MRIB.

See the Advanced Configuration Guide for more information about IMPM as well as detailed configuration examples.

IMPM allows the system to manage Layer 2 and Layer 3 IP multicast flows by sorting them into the available multicast paths through the switch fabric. The ingress multicast manager tracks the amount of available multicast bandwidth per path and the amount of bandwidth used per IP multicast stream. The following traffic is managed by IMPM when enabled:

• IPv4 and IPv6 routed multicast traffic

• VPLS IGMP snooping traffic

• VPLS PIM snooping for IPv4 traffic

• VPLS PIM snooping for IPv6 traffic (only when sg-based forwarding is configured)

Two policies define how each path should be managed, the bandwidth policy, and how multicast channels compete for the available bandwidth, the multicast information policy.

Chassis multicast planes should not be confused with IOM/IMM multicast paths. The IOM/IMM uses multicast paths to reach multicast planes on the switch fabric. An IOM/IMM may have less or more multicast paths than the number of multicast planes available in the chassis.

Each IOM/IMM multicast path is either a primary or secondary path type. The path type indicates the multicast scheduling priority within the switch fabric. Multicast flows sent on primary paths are scheduled at multicast high priority while secondary paths are associated with multicast low priority.

The system determines the number of primary and secondary paths from each IOM/IMM forwarding plane and distributes them as equally as possible between the available switch fabric multicast planes. Each multicast plane may terminate multiple paths of both the primary and secondary types.

The system ingress multicast management module evaluates the ingress multicast flows from each ingress forwarding plane and determines the best multicast path for the flow. A specific path can be used until the terminating multicast plane is ‟maxed” out (based on the rate limit defined in the per-mcast-plane-capacity commands) at which time either flows are moved to other paths or potentially blackholed (flows with the lowest preference are dropped first). In this way, the system makes the best use of the available multicast capacity without congesting individual multicast planes.

The switch fabric is simultaneously handling both unicast and multicast flows. The switch fabric uses a weighted scheduling scheme between multicast high, unicast high, multicast low and unicast low when deciding which cell to forward to the egress forwarding plane next. The weighted mechanism allows some amount of unicast and lower priority multicast (secondary) to drain on the egress switch fabric links used by each multicast plane. The amount is variable based on the number of switch fabric planes available on the amount of traffic attempting to use the fabric planes. The per-mcast-plane-capacity commands allows the amount of managed multicast traffic to be tuned to compensate for the expected available egress multicast bandwidth per multicast plane. In conditions where it is highly desirable to prevent multicast plane congestion, the per-mcast-plane-capacity commands should be used to compensate for the non-multicast or secondary multicast switch fabric traffic.

For most Layer 2 switches, multicast traffic is treated like an unknown MAC address or broadcast frame, which causes the incoming frame to be flooded out (broadcast) on every port within a VLAN. While this is acceptable behavior for unknowns and broadcasts, as IP Multicast hosts may join and be interested in only specific multicast groups, all this flooded traffic results in wasted bandwidth on network segments and end stations.

IGMP snooping entails using information in layer 3 protocol headers of multicast control messages to determine the processing at layer 2. By doing so, an IGMP snooping switch provides the benefit of conserving bandwidth on those segments of the network where no node has expressed interest in receiving packets addressed to the group address.

On the 7450 ESS and 7750 SR, IGMP snooping can be enabled in the context of VPLS services. The IGMP snooping feature allows for optimization of the multicast data flow for a group within a service to only those Service Access Points (SAPs) and Service Distribution Points (SDPs) that are members of the group. In fact, the 7450 ESS and 7750 SR implementation performs more than pure snooping of IGMP data, because it also summarizes upstream IGMP reports and responds to downstream queries.

The 7450 ESS and 7750 SR maintain several multicast databases:

• A port database on each SAP and SDP lists the multicast groups that are active on this SAP or SDP.

• All port databases are compiled into a central proxy database. Towards the multicast routers, summarized group membership reports are sent based on the information in the proxy database.

• The information in the different port databases is also used to compile the multicast forwarding information base (MFIB). This contains the active SAPs and SDPs for every combination of source router and group address (S,G), and is used for the actual multicast replication and forwarding.

When the router receives a join report from a host for a multicast group, it adds the group to the port database and (if it is a new group) to the proxy database. It also adds the SAP or SDP to existing (S,G) in the MFIB, or builds a new MFIB entry.

When the router receives a leave report from a host, it first checks if other devices on the SAP or SDP still want to receive the group (unless fast leave is enabled). Then it removes the group from the port database, and from the proxy database if it was the only receiver of the group. The router also deletes entries if it does not receive periodic membership confirmations from the hosts.

The fast leave feature finds its use in multicast TV delivery systems, for example. Fast Leave speeds up the membership leave process by terminating the multicast session immediately, instead of the standard procedure of issuing a group specific query to check if other group members are present on the SAP or SDP.

Multicast VPLS Registration (MVR) is a bandwidth optimization method for multicast in a broadband services network. MVR allows a subscriber on a port to subscribe and unsubscribe to a multicast stream on one or more network-wide multicast VPLS instances.

MVR assumes that subscribers join and leave multicast streams by sending IGMP join and leave messages. The IGMP leave and join message are sent inside the VPLS to which the subscriber port is assigned. The multicast VPLS is shared in the network while the subscribers remain in separate VPLS services. Using MVR, users on different VPLS cannot exchange any information between them, but still multicast services are provided.

On the MVR VPLS, IGMP snooping must be enabled. On the user VPLS, IGMP snooping and MVR work independently. If IGMP snooping and MVR are both enabled, MVR reacts only to join and leave messages from multicast groups configured under MVR. Join and leave messages from all other multicast groups are managed by IGMP snooping in the local VPLS. This way, potentially several MVR VPLS instances could be configured, each with its own set of multicast channels.

MVR by proxy — In some situations, the multicast traffic should not be copied from the MVR VPLS to the SAP on which the IGMP message was received (standard MVR behavior) but to another SAP. This is called MVR by proxy.

MVR and MVR by proxy shows a MVR and MVR by proxy configuration.

Layer 3 multicast load balancing establishes a more efficient distribution of Layer 3 multicast data over ECMP and LAG links. Operators have the option to redistribute multicast groups over ECMP or LAG links if the number of links changes either up or down.

When implementing this feature, there are several considerations. When multicast load balancing is not configured, the distribution remains as is. Multicast load balancing is based on the number of ‟s,g” groups. This means that bandwidth considerations are not considered. The multicast groups are distributed over the available links as joins are processed. When link failure occurs, the load is distributed on the failed channel to the remaining channels so multicast groups are evenly distributed over the remaining links. When a link is added (or failed link returned) all multicast joins on the added links are allocated until a balance is achieved.

When multicast load balancing is configured, but the channels are not found in the multicast-info-policy, then multicast load balancing is based on the number of ‟s,g” groups. This means that bandwidth considerations are not considered. The multicast groups are distributed over the available links as joins are processed. The multicast groups are evenly distributed over the remaining links. When link failure occurs, the load is distributed on the failed channel to the remaining channels. When a link is added (or failed link returned) all multicast joins on the added links are allocated until a balance is achieved. A manual redistribute command enables the operator to re-evaluate the current balance and, if required, move channels to different links to achieve a balance. A timed redistribute parameter allows the system to automatically, at regular intervals, redistribute multicast groups over available links. If no links have been added or removed from the ECMP/LAG interface, then no redistribution is attempted.

When multicast load balancing is configured, multicast groups are distributed over the available links as joins are processed based on bandwidth configured for the specified group address. If the bandwidth is not configured for the multicast stream then the configured default value is used.

If link failure occurs, the load is distributed on the failed channel to the remaining channels. The bandwidth required over each individual link is evenly distributed over the remaining links.

When an additional link is available for a specific multicast stream, then it is considered in all multicast stream additions applied to the interface. This multicast stream is included in the next scheduled automatic rebalance run. A rebalance run re-evaluates the current balance with regard to the bandwidth utilization and if required, move multicast streams to different links to achieve a balance.

A rebalance, either timed or executing the mc-ecmp-rebalance command, should be administered gradually to minimize the effect of the rebalancing process on the different multicast streams. If multicast re-balancing is disabled and subsequently enabled, keeping with the rebalance process, the gradual and least invasive method is used to minimize the effect of the changes to the customer.

By default multicast load balancing over ECMP links is enabled and set at 30 minutes.

The rebalance process can be executed as a low priority background task while control of the console is returned to the operator. When multicast load rebalancing is not enabled, then ECMP changes are not be optimized, however, when a link is added occurs an attempt is made to balance the number of multicast streams on the available ECMP links. This however may not result in balanced utilization of ECMP links.

Only a single mc-ecmp-rebalance command can be executed any specific time, if a rebalance is in progress and the command is entered, it is rejected with the message saying that a rebalance is already in progress. A low priority event is generated when an actual change for a specific multicast stream occurs as a result of the rebalance process.

The target application for this feature is linear TV delivery. In some countries, wholesale Service Providers are obligated by the government regulation to provide information about channel viewership per subscriber to retailers.

A service provider (wholesaler or retailer) my use this information for:

• billing purposes

• market research/data mining to gain view into the most frequently watched channels, duration of the channel viewing, frequency of channel zapping by the time of the day, and so on

The information about channel viewership is based on IGMP states maintained per each subscriber host. Each event related to the IGMP state creation is recorded and formatted by the IGMP process. The formatted event is then sent to another task in the system (Exporter), which allocates a TX buffer and start a timer.

The event is then be written by the Exporter into the buffer. The buffer in essence corresponds to the packet that contains a single event or a set of events. Those events are transported as data records over UDP transport to an external collector node. The packet itself has a header followed by a set of TLV type data structures, each describing a unique filed within the IGMP event.

The packet is transmitted when it reaches a preconfigured size (1400 bytes), or when the timer expires, whichever comes first.

The receiving end (collector node) accepts the data on the destination UDP port. It must be aware of the data format so that it can interpret incoming data accordingly. The implementation details of the receiving node are outside of the scope of this description and are left to the network operator.

The IGMP state recording per subscriber host must be supported for hosts which are replicating multicast traffic directly as well as for those host that are only keeping track of IGMP states for the HQoS Adjustment purpose. The latter is implemented by redirection and not the Host Tracking (HT) feature as originally proposed. The IGMP reporting must differentiate events between direct replication and redirection.

It further distinguish events that are related to denial of IGMP state creation (because of filters, MCAC failure, and so on) and the ones that are related to removal of an already existing IGMP state in the system.

There are several deployment scenarios for delivering multicast directly over subscriber hosts:

• 1:1 model (subscriber per VLAN/SAP) with the Access Node (AN) that is not IGMP/MLD aware.

• N:1 model (service per VLAN/SAP) with the AN in the Snooping mode.

• N:1 model with the AN in the Proxy mode.

• N:1 model with the AN that is not IGMP/MLD aware.

There are two modes of operation for subscriber multicast that can be chosen to address the above mentioned deployment scenarios:

1. Per SAP replication — A single multicast stream per group is forwarded on any given SAP. Even if the SAP has a multicast group (channel) that is registered to multiple hosts, only a single copy of the multicast stream is forwarded over this SAP. The multicast stream has a multicast destination MAC address (as opposed to unicast). IGMP/MLD states are maintained per host. This is the default mode of operation.

2. Per subscriber host replication in this mode of operation, multicast is replicated per subscriber host even if this means that multiple copies of the same stream that are forwarded over the same SAP. For example, if two hosts on the same SAP are registered to receive the same multicast group (channel), then this multicast channel is replicated twice on the same SAP. The streams have a unique unicast destination MAC address (otherwise it would not make sense to replicate the streams twice).

In all deployment scenarios and modes of operation the IGMP/MLD states per source IP address of the incoming IGMP/MLD message is maintained. This source IP address might represent a subscriber hosts or the AN (proxy mode).

For MLD, the source IP address is the host link local address. Therefore, the MLD message is associated with all IP address/prefix of the host.

In a PPPoE environment, multicast replication is performed per session (host) regardless of whether those sessions are shared per SAP or they reside on individual SAPs. This is because of the point-to-point nature of PPPoE sessions. HQoS adjustment is not needed as multicast is part of the PPPoE session traffic that is flowing through subscriber queues. Multicast packets are sent with unicast MAC address to each CPE. PPP protocol field is set to IP and the destination IP address is the multicast group address for each unique session ID.

This model is shown in Multicast IPv4 address and unicast MAC address in PPPoE subscriber multicast.

Note that in a SRRP setup, the multi-chassis active BNG (SRRP master state) uses the SRRP MAC address as the source MAC for both multicast and multicast control protocol packets. This is only applicable to PPPoE hosts.

The query function in IGMP can cause some unintended flooding in N:1 IPoE deployment model with AN in the IGMP snooping mode. By maintaining IGMP session states per host, it is assumed that the IGMP interaction between multicast receivers and the BNG are on a one-to-one basis. Upon arrival of an IGMP leave from a host for a specific multicast group, the IGMP querier would normally multicast a group-specific query (fast-leave). In N:1 model with SAP replication mode enabled, the 7750 SR and 7450 ESS sends a group-specific query (fast-leave) only when it receives the IGMP leave message for the last group shared amongst all subscribers on this SAP.

IGMP/MLD timers are maintained under the following hierarchy:

IPv4:

config>router>igmp

config>service>vprn>igmp

IPv6:

config>router>mld

config>service>vprn>mld

The IGMP/MLD timers are controlled on a per routing instance (VRF or GRT) level.

The timer values are used to:

• determine the interval at which queries are transmitted (query-interval)

• determine the amount of time after which a join times out

However, the timers can be different for hosts and redirected interface in case that redirection between VRFs is enabled.

IGMP/MLD query related intervals (query-interval, query-last-member-interval, query-response-interval, robust-count) are configured on a global router/vprn IGMP/MLD level. They are used to determine the IGMP/MLD timeout states and the rates at which queries are transmitted.

In case of redirection, the subscriber-host IGMP/MLD state determines the IGMP/MLD state on the redirected interface, assuming that IGMP/MLD messages are not directly received on the redirected interface (for example from the AN performing IGMP/MLD forking). For example, if the redirected interface is not receiving IGMP/MLD messages from the downstream node, then the IGMP/MLD state under the redirected interface is removed simultaneously with the removal of the IGMP/MLD state for the subscriber host (because of leave or a timeout).

If the redirected interface is receiving IGMP/MLD message directly from the downstream node, the IGMP/MLD states on that redirected interface are driven by those direct IGMP/MLD messages.

For example, an IGMP/MLD host in VRF1 has an expiry time of 60 seconds and the expiry time defined under the VRF2 where multicast traffic is redirected is set to 90 seconds. The IGMP/MLD state times out for the host in VRF1 after 60s, and if no host has joined the same multicast group in VRF2 (where redirected interface resides), the IGMP state is removed there too.

If a join was received directly on the redirection interface in VRF2, the IGMP/MLD state for that group is maintained for 90s, regardless of the IGMP/MLD state for the same group in VRF1.

HQoS Adjustment is required in the scenarios where subscriber multicast traffic flow is disassociated from subscriber queues. In other words, the unicast traffic for the subscriber is flowing through the subscriber queues while at the same time multicast traffic for the same subscriber is explicitly (through redirection) or implicitly (per-sap replication mode) redirected through a separate non-subscriber queue. In this case HQoS Adjustment can be deployed where preconfigured multicast bandwidth per channel is artificially included in HQoS. For example, bandwidth consumption per multicast group must be known in advance and configured within the 7450 ESS and 7750 SR. By keeping the IGMP state per host, the bandwidth for the multicast group (channel) to which the host is registered is known and is deducted as consumed from the aggregate subscriber bandwidth.

The multicast bandwidth per channel must be known (this is always an approximation) and provisioned in the BNG node in advance.

In PPPoE and in IPoE per-host replication environment, HQoS Adjustment is not needed as multicast traffic is unicast to each subscriber and therefore is flowing through subscriber queues.

For HQoS Adjustment, the channel bandwidth definition and association with an interface is the same as in the MCAC case. This is a departure from the legacy HT channel bandwidth definition which is done by a multicast-info-policy.

Example of HQoS adjustment:

Channel definition:

configure
    router 
        mcac
            policy <name>
                <channel definition>

Channel bandwidth definition policy can be applied under:

• group-interface

  configure
      service vprn <id>
          igmp
              group-interface <ip-int-name>
                  mcac
                      policy <mcac-policy-name>  
  

• plain interface

  configure
      router/service vprn
          igmp
              interface <ip-int-name>
                  mcac
                      policy <mcac-policy-name>
  

• retailer group-interface

  configure 
      service vprn <id>
          igmp
              group-interface fwd-service <svc-id> <grp-if-name>  
                  mcac
    policy <mcac-policy-name>
  

Enabling HQoS adjustment:

configure
    subscriber-management
        igmp-policy <name>
            egress-rate-modify [egress-aggregate-rate-limit | scheduler <name>]

Applying HQoS adjustment to the subscriber:

configure
    subscriber-management
        sub-profile <subscriber-profile-name>
            igmp-policy <name>

To activate HQoS adjustment on the subscriber level, the sub-mcac-policy must be enabled under the subscriber with the following CLI:

configure
    subscriber-management
        sub-mcac-policy <pol-name>
            no shutdown

configure
    subscriber-management
        sub-profile <subscriber-profile-name>
            sub-mcac-policy <pol-name> 

The adjusted bandwidth during operation can be verified with the following commands (depending whether agg-rate-limit or scheduler-policy is used):

*B:BNG-1# show service active-subscribers subscriber "sub-1" detail  
===============================================================================
Active Subscribers
===============================================================================
-------------------------------------------------------------------------------
Subscriber sub-1 
-------------------------------------------------------------------------------
I. Sched. Policy : up-silver             
E. Sched. Policy : N/A                              E. Agg Rate Limit: 4000
I. Policer Ctrl. : N/A                              
E. Policer Ctrl. : N/A                              
Q Frame-Based Ac*: Disabled                         
Acct. Policy     : N/A                              Collect Stats    : Enabled
Rad. Acct. Pol.  : sub-1-acct                         
Dupl. Acct. Pol. : N/A                              
ANCP Pol.        : N/A                              
HostTrk Pol.     : N/A                              
IGMP Policy      : sub-1-IGMP-Pol                     
Sub. MCAC Policy : sub-1-MCAC                          
NAT Policy       : N/A                              
Def. Encap Offset: none                             Encap Offset Mode: none
Avg Frame Size   : N/A                              
Preference       : 5                                
Sub. ANCP-String : "sub-1"
Sub. Int Dest Id : ""
Igmp Rate Adj    : -2000                            
RADIUS Rate-Limit: N/A                              
Oper-Rate-Limit  : 2000
...
-------------------------------------------------------------------------------
*B:BNG-1#

Consider a different example with a scheduler instead of agg-rate-limit:

*A:Dut-C>config>subscr-mgmt>sub-prof# info 
----------------------------------------------
            igmp-policy "pol1"
            sub-mcac-policy "smp"
            egress
                scheduler-policy "h1"
                    scheduler "t2" rate 30000  
                exit
            exit
----------------------------------------------

*A:Dut-C>config>subscr-mgmt>igmp-policy# info 
----------------------------------------------
            egress-rate-modify scheduler "t2"
            redirection-policy "mc_redir1"
----------------------------------------------

Assume that the subscriber joins a new channel with bandwidth of 1 Mb/s (1000 kb/s).

A:Dut-C>config>subscr-mgmt>sub-prof>egr>sched# show qos scheduler-hierarchy subscrib
er "sub_1" detail 
===============================================================================
Scheduler Hierarchy - Subscriber sub_1
===============================================================================
Ingress Scheduler Policy: 
Egress Scheduler Policy : h1
-------------------------------------------------------------------------------
Legend :
(*) real-time dynamic value
(w) Wire rates
B   Bytes
-------------------------------------------------------------------------------
Root (Ing)
|
No Active Members Found on slot 1

Root (Egr)
| slot(1)
|--(S) : t1  
|   |    AdminPIR:90000      AdminCIR:10000     
|   |
|   |
|   |    [Within CIR Level 0 Weight 0]
|   |    Assigned:0          Offered:0         
|   |    Consumed:0         
|   |
|   |    [Above CIR Level 0 Weight 0]
|   |    Assigned:0          Offered:0         
|   |    Consumed:0         
|   |    TotalConsumed:0         
|   |    OperPIR:90000     
|   |
|   |    [As Parent]
|   |    Rate:90000     
|   |    ConsumedByChildren:0         
|   |                                 
|   |
|   |--(S) : t2  
|   |   |    AdminPIR:29000      AdminCIR:10000(sum)       <==== bw 1000 from igmp s
ubstracted
|   |   |
|   |   |
|   |   |    [Within CIR Level 0 Weight 1]
|   |   |    Assigned:10000      Offered:0         
|   |   |    Consumed:0         
|   |   |
|   |   |    [Above CIR Level 1 Weight 1]
|   |   |    Assigned:29000      Offered:0                 <==== bw 1000 from igmp s
ubstracted    
|   |   |    Consumed:0         
|   |   |
|   |   |
|   |   |    TotalConsumed:0         
|   |   |    OperPIR:29000                                 <==== bw 1000 from igmp s
ubstracted
|   |   |
|   |   |    [As Parent]
|   |   |    Rate:29000                                     <==== bw 1000 from igmp 
substracted
|   |   |    ConsumedByChildren:0         
|   |   |
|   |   |
|   |   |--(S) : t3  
|   |   |   |    AdminPIR:70000      AdminCIR:10000     
|   |   |   |
|   |   |   |
|   |   |   |    [Within CIR Level 0 Weight 1]
|   |   |   |    Assigned:10000      Offered:0         
|   |   |   |    Consumed:0         
|   |   |   |
|   |   |   |    [Above CIR Level 1 Weight 1]
|   |   |   |    Assigned:29000      Offered:0         
|   |   |   |    Consumed:0         
|   |   |   |
|   |   |   |
|   |   |   |    TotalConsumed:0         
|   |   |   |    OperPIR:29000     
|   |   |   |
|   |   |   |    [As Parent]
|   |   |   |    Rate:29000           
|   |   |   |    ConsumedByChildren:0         
|   |   |   |

*A:Dut-C>config>subscr-mgmt>igmp-policy# show service active-subscribers sub-mcac 
===============================================================================
Active Subscribers Sub-MCAC
===============================================================================
Subscriber                             : sub_1
MCAC-policy                            : smp (inService)
In use mandatory bandwidth             : 1000
In use optional bandwidth              : 0
Available mandatory bandwidth          : 1147482647
Available optional bandwidth           : 1000000000
-------------------------------------------------------------------------------
Subscriber                             : sub_2
MCAC-policy                            : smp (inService)
In use mandatory bandwidth             : 0
In use optional bandwidth              : 0
Available mandatory bandwidth          : 1147483647
Available optional bandwidth           : 1000000000
-------------------------------------------------------------------------------
-------------------------------------------------------------------------------
Number of Subscribers : 2
===============================================================================
*A:Dut-C# 

*A:Dut-C# show service active-subscribers subscriber "sub_1" detail 
===============================================================================
Active Subscribers
===============================================================================
-------------------------------------------------------------------------------
Subscriber sub_1 (1)
-------------------------------------------------------------------------------
I. Sched. Policy : N/A                              
E. Sched. Policy : h1                               E. Agg Rate Limit: Max
I. Policer Ctrl. : N/A                              
E. Policer Ctrl. : N/A                              
Q Frame-Based Ac*: Disabled                         
Acct. Policy     : N/A                              Collect Stats    : Disabled
Rad. Acct. Pol.  : N/A                              
Dupl. Acct. Pol. : N/A                              
ANCP Pol.        : N/A                              
HostTrk Pol.     : N/A                              
IGMP Policy      : pol1                             
Sub. MCAC Policy : smp                              
NAT Policy       : N/A                              
Def. Encap Offset: none                             Encap Offset Mode: none
Avg Frame Size   : N/A                              
Preference       : 5                                
Sub. ANCP-String : "sub_1"
Sub. Int Dest Id : ""
Igmp Rate Adj    : N/A                              
RADIUS Rate-Limit: N/A                              
Oper-Rate-Limit  : Maximum                          
...
===============================================================================
*A:Dut-C# 


*A:Dut-C# show subscriber-mgmt igmp-policy "pol1" 
===============================================================================
IGMP Policy pol1
===============================================================================
Import Policy                         : 
Admin Version                         : 3
Num Subscribers                       : 2
Host Max Group                        : No Limit
Host Max Sources                      : No Limit
Fast Leave                            : yes
Redirection Policy                    : mc_redir1
Per Host Replication                  : no
Egress Rate Modify                    : "t2"
Mcast Reporting Destination Name      : 
Mcast Reporting Admin State           : Disabled
===============================================================================
*A:Dut-C# 

Two levels of MCAC can be enabled simultaneously and in such case this is referred as Hierarchical MCAC (H-MCAC). In case that redirection is enabled, H-MCAC per subscriber and the redirected interface is supported. However, mcac per group-interface in this case is not supported. Channel definition policy for the subscriber and the redirected interface is in this case referenced under the igmp>interface (redirected interface) CLI or for IPv6 mld>interface.

If redirection is disabled, H-MCAC for both, the subscriber and the group-interface is supported. The channel definition policy is in this case configured under the config>router>igmp>group interface context or for IPv6 config>router>mld>group interface.

There are two options in multicast redirection. The first option is to redirect all subscriber multicast traffic to a dedicated redirect interface.

Example:

Defining redirection action:

configure 
    router
        policy-options
            begin 
            policy-statement <name>
                default-action accept 
                multicast-redirection [fwd-service <svc id>] <interface name>
                exit
            exit
        exit
    exit

The second option is to redirect only specific multicast groups to the redirect interface while the remaining groups remains on the subscriber SAP. This is applicable for both IPv4 and IPv6. For IPv6 host-ip for a policy statement is not supported.

Example:

Defining redirection action:

configure 
    router
        policy-options
            begin 
            prefix-list <name>
                prefix <IPv4 multicast groups>
                prefix <IPv4 multicast groups
            exit
            policy-statement <name>
                entry 1
                    from
                        group-address <prefix-list name>
                    action accept
                    multicast-redirection [fwd-service <svc id>] <interface name>
                exit
            exit
    exit

Applying redirection to the subscriber for IGMP and MLD respectively.

    configure
        subscr-mgmt
            igmp-policy <name>
                redirection-policy <name>
                exit
            exit
            mld-policy <name>
                redirection-policy <name>

Redirection that cross-connects GRT and VPRN is not supported. Redirection can be only performed between interfaces in the GRT, or between the interfaces in any of the VPRN (cross connecting VPRNs is allowed).

Redirection is also supported in a wholesaler/retailer VPRN model where redirected Layer 3 interface resides in the retailer VPRN.

MCAC is supported on three levels:

• per subscriber

• per group-interface

• per redirected interface

MCAC is supported for the following replication modes:

• per-host

• per-SAP

• per-SPI

• redirected interface

Two levels of MCAC can be enabled simultaneously and in such case this is referred as Hierarchical MCAC (H-MCAC). If redirection is enabled, H-MCAC per subscriber and the redirected interface is supported. However, MCAC per group-interface in this case is not supported. Channel definition policy for the subscriber and the redirected interface is in this case referenced under the config>router>igmp>interface (redirected interface) CLI hierarchy or IPv6 config>router>mld>interface.

If redirection is disabled, H-MCAC for both, the subscriber and the group-interface is supported. The channel definition policy is in this case configured under the config>router>igmp>group-interface CLI hierarchy or IPv6 config>router>mld>group-interface.

Examples

Channel definition:

configure
    router 
        mcac    
            policy <mcac-pol-name>
                bundle <bundle-name>
                    bandwidth <kbps>
                    channel <start-address> <end-address> bw <bw> [class {high|low}]
 [type {mandatory|optional}]
                    :
                    :

Channel bandwidth definition policy can be referenced under the:

• group-interface (this is used for subscribers when redirection is disabled)

  configure
                   service vprn <id>
              igmp/mld    
                   group-interface <ip-int-name>
                      mcac
                          policy <mcac-policy-name> 
                          policy <mcac-policy-name> 
  

• plain interface

  configure
      router
          igmp/mld
              interface <ip-int-name>
                  mcac
                      policy <mcac-policy-name>
  
  configure
      service vprn <id>
          igmp/mld
              interface <if-name>
                  mcac
                      unconstrained-bw <bandwidth> mandatory-bw <mandatory-bw>
  
  

• retailer’s VPRN reference the group-interface in the wholesaler’s VPRN

  configure 
      service vprn <id>
          igmp/mld                                
              group-interface fwd-service <svc-id> <ip-int-name>  
                  mcac
                      policy <mcac-policy-name>
  
  

Enabling MCAC per subscriber:

configure
    subscr-mgmt
        sub-mcac-policy <name>
            unconstrained-bw <bandwidth> mandatory-bw <mandatory-bw>
  
configure
    subscr-mgmt
        sub-profile <subscriber-profile-name>
            sub-mcac-policy <name>

Enabling MCAC per-group-interface

configure
        service vprn <id>
            igmp/mld        
                group-interface <ip-int-name>
                    mcac
                        unconstrained-bw <bandwidth> mandatory-bw <mandatory-bw>    

Enabling MCAC per redirected interface

configure
    router
        igmp/mld    
            interface <ip-int-name>
                mcac
                    unconstrained-bw <bandwidth> mandatory-bw <mandatory-bw>

configure
    service vprn <id>
        igmp/mld    
            interface <ip-int-name>
                mcac
                    unconstrained-bw <bandwidth> mandatory-bw <mandatory-bw>

Channel bandwidth definition (by a MCAC policy) can be applied under the interface level (group-interface or regular interface):

config>router/service>igmp/mld>interface/grp-if

The following configuration options can lead to the confusion as to which MCAC policy is in effect:

• The MCAC policy (channel bandwidth definition) can be applied under two different places (grp-if or regular intf).

• The same policy is used for (H)MCAC and HQoS Adjust with redirection enabled or disabled.

The general, the rule is that the MCAC policy under the group-interface is always be in effect in cases where redirection is disabled. This is valid for subscriber or group-interface MCAC, H-MCAC (subscriber and group-interface) and HQoS Adjust in per SAP replication mode.

If redirection is enabled, the MCAC policy under the group-interface is ignored.

If redirection is enabled, but there is no MCAC policy applied under the redirected interface (redirected interface is the interface to which IGMP/MLD Joins are redirected from subscriber hosts), regardless of whether the MCAC policy under the group-interface is applied or not, then:

• HQoS adjust has no effect.

• MCAC has no effect not only per redirected interface but also per subscriber.

Multicast filtering must be done per session (host) for IPoE and PPPoE. There are two types of filters that are supported:

• IGMP filters on access ingress. Those filters control the flow of IGMP messages between the host and the BNG. They are applied with the import statement in the igmp-policy. The same filters are used for multicast-redirection policy:

  For IPv4:

  configure
      subscr-mgmt
          igmp-policy <name>
              import <policy-name>
  

  For IPv6:

  configure
      subscr-mgmt
          mld-policy <name>
              import <policy-name>
  

  An example of the filter definition is shown below:

  configure
      router
          policy-options
              begin
              prefix-list <pref-name>
                  prefix <pref-definition>
              policy-statement <name>
                  entry 1
                      from
                          group-address <pref-name>
                          source-address <ip>
                          protocol igmp
                      exit
                      action accept
                      exit
                  exit
                  default-action reject
  

• Regular traffic filters where control multicast traffic flow can be controlled in both directions (ingress/egress). This is supported through ip-filters under the SLA profile.

For IPv6 specifically, up to 15 import policies can be applied to the subscriber IPv6 host. Each of the import policies can be configured as an IPTV package. Depending on the IPTV package the subscriber ordered, a list of import policies is applied at authentication.

At authentication, the first 14 import policies are applied either through LUDB or RADIUS using a list of VSAs. These 14 import policies are not read in any particular order and to make the list deterministic as a black list or a white list, the import policy inside the MLD policy must be configured. The import policy inside the MLD policy is always applied as the last policy which determines whether the list of 14 policies is a white list or a black list. For example, to make the list of 15 import policies a white list, the first 14 policies should have action configured as accept and the last import policy inside the MLD policy should have default action configured as reject. For a black list, the first 14 policies should have action configured as reject and the last import policy inside the MLD policy should have default action configured as accept. It is recommended to configure the MLD import policy with a default action without an import list which is used to determine if the list is a black or white list. If the last import policy is overridden, it changes the MLD policy which in turn overrides the subscriber profile.

It is possible to make a mixed black and white list, where the general rule is that the import policy inside the MLD policy is always the last to be applied to the subscriber.

The list of MLD import policies can be overridden either through RADIUS CoA or through the tools>perform>subscriber-management>coa command. There are two VSAs for overriding the list of import policies. The first VSA, Alc-Mld-Import-Policy, is used to completely override the first 14 MLD policies. A RADIUS CoA can contain up to 14 VSAs to override the existing list. Because of limitations with the length of a CLI command, up to five VSAs can be passed to override the existing list when using the tools command. The second VSA, Alc-Mld-Import-Policy-Modif, allows the addition or subtraction of a single MLD import policy. To add an import policy to the existing list, use a prefix of ‛a:’ and to remove an import policy, use a prefix of ‛s:’, for example, ‛a:sport-pack1-hd’ or ‛s:movie-pack1-hd’. Multiple import policies can be added or removed within a single CoA by using multiple VSAs. Up to five import policies can be added or removed using a single CoA. The last import policy inside the MLD policy can be overridden using the subscriber profile.

If a subscriber is not associated with an MLD policy, the import policies applied either at authentication or overridden in mid-session are always stored while the subscriber is online. The import policy does not take effect as long as the subscriber is not associated with an MLD policy. However, as soon as the subscriber is associated with an MLD policy, the list of import policies stored is instantly applied.

The delivery of multicast to the subscribers-interface in a VPRN environment depends on the multicast deployment model (PIM, mBGP). In each model, a subscriber-interface is treated as a regular CE-PE interface that has registered v4 multicast listeners.

Multicast support on subscriber interfaces is supported in both wholesale/retail models:

• wholesale/retail VPRN (IPoE and PPPoE)

• LAC/LNS (PPPoE only)

  This model is shown in Wholesale/retail multicast support.

  

The distinction between these two models is that in the case of LAC/LNS, the replication is performed further up in the network on an LNS node. This means that the traffic between LAC and LNS is multiplied by the amount of replications.

In per-sap replication mode (which is applicable only to IPoE subscribers), multicast traffic is forwarded through the SAP queue which is outside of the subscriber’s queues and therefore not accounted in subscriber aggregate rate limit. HQoS Adjust is used to remedy this situation.

If the SAP queue is removed from the static SAP in IPoE 1:1 model (with profiled-only-traffic command), multicast traffic flows through internal queues which cannot be tied into a port-scheduler as part of HQoS. Consequently, the port-scheduler max-rate as defined in the port scheduler policy is used only to rate limit unicast traffic. In other words, the max-rate value in the port scheduler policy must be lowered for the amount of anticipated multicast traffic that flows through the port where the port scheduler policy is applied.

A similar logic applies to the per-sap replication mode on dynamic SAPs (MSAPs) even if the SAP queue is not removed. Although the multicast traffic is flowing through the SAP queue in this case, the SAP QoS policy on MSAP cannot be changed from the default one. The default QoS policy on a SAP contains a single queue that is not parented by the port-scheduler.

Those restrictions do not apply to static SAPs where the SAP QoS policy can be customized and its queues consequently tied to the port-scheduler.

The subscriber can receive multicast content through the subscriber SAP, the redirected interface, or a combination of both.

Multicast redundancy is only supported on a MC-LAG topology. Multicast traffic can be delivered over the subscriber SAP, the redirected interface, or a combination of both.

Subscriber IGMP states can be synchronized across multiple routers to ensure minimal interruption of (video delivery) service during network outages. The IGMP/MLD state of a subscriber-host in the system is tied to the state of the underlying MC-LAG protection mechanism. For example, IGMP states is activated only for subscribers that are anchored under the group-interfaces with master SRRP state or under an active MC-LAG port.

For multicast redirection on a MC-LAG topology, it must be ensured that the redirected interface (the interface to which multicast forwarding is redirected) is under the same MC-LAG as the subscriber. Otherwise, IGMP states on the redirected interface is derived independently of the IGMP states for the subscriber from which IGMP/MLD messages are redirected.

The IGMP/MLD synchronization process in conjunction with underlying access protection mechanisms works as follows:

IGMP/MLD states for the subscriber updates only if IGMP/MLD messages (Joins/Leaves/Reports, and so on) are received directly from the downstream node on an active MC-LAG link. This is valid irrespective of the IGMP querier status for the subscriber.

In all other cases, assuming that some protection mechanism in the access is present, the IGMP/MLD messages are discarded and consequently no IGMP/MLD state is updated. Similar logic applies to regular Layer 3 interfaces, where SRRP is replaced with VRRP.

• After the subscriber IGMP/MLD state is updated as a result of a directly-received IGMP/MLD message on an active subscriber (SRRP master state or active MC-LAG), the sync IGMP/MLD message is sent to the standby subscriber over the Multi-Chassis Synchronization (MCS) protocol. The synchronized IGMP state is populated in the MCS database in all pairing routers.

• If an IGMP/MLD sync (MCS) message is received from the peering node, the IGMP state for the standby subscriber is updated in the MCS database but it is not downloaded into the forwarding plane unless there is a switchover. If the IGMP/MLD sync message is received for the active subscriber, the message is discarded.

• IGMP/MLD queries are sent out only by IGMP/MLD querier. In an MC-LAG environment, this is the node with the active MC-LAG link.

  Note: MC-LAG is usually configured with SRRP and the SRRP state is derived from the MC-LAG.

• IGMP/MLD states from the MCS database are:

  • activated on non-querier subscriber in case that neither SRRP nor MC-LAG is deployed. It is assumed that the querier subscriber has received the original IGMP/MLD message and consequently sent the IGMP/MLD MCS Sync to the non-querier (standby). Non-querier interface accepts the MCS sync message and also it propagate these IGMP/MLD states to PIM.

    The querier subscriber does not accept the IGMP/MLD update from the MCS database.

  • aware of the state of MC-LAG. As soon as the standby MC-LAG becomes active, the IGMP/MLD states is activated and is propagated to PIM. Traffic is forwarded as soon as multicast streams are delivered to the node and the IGMP states under the subscriber are activated. On a standby MC-LAG, IGMP states are not propagated from the MCS database to PIM and consequently subscribers.

  • aware of the SRRP state. Because the subscriber with SRRP master state is considered active, the states are propagated to PIM as well. On standby SRRP, IGMP states are be propagated from MCS database to PIM and consequently to subscribers.

• For MC-lag setups, after the switchover is triggered by MC-LAG or SRRP, the IGMP/MLD states from MCS database on the newly active MC-LAG node or subscriber under the new SRRP instance in the master state is sent to PIM and consequently to the forwarding plane effectively turning on multicast forwarding.

An active and standby subscriber refers to the state of underlying protection mechanism (active MC-LAG).

To summarize, in multi-chassis environment with subscribers, IGMP synchronization enabled and an access layer protection mechanism in place (MC-LAG), the behavior for is the following:

• IGMP/MLD states are synchronized between the chassis.

• On a MC-LAG setup, only the SRRP instance in master state or active MC-LAG forwards downstream multicast traffic.

• Length of outage during the switchover is determined by the detection and recovery of the underlying protection mechanism (MC-LAG or MCS) in addition to local propagation of IGMP/MLD states from MCS database to PIM and consequently to forwarding plane.

  Note: IGMP/MLD states can be statically configured on both redundant nodes to attract multicast traffic from upstream and therefore minimize outage during the switchover.

The query interval commands—query-interval, query-last-listener-interval, query-last-member-interval, and query-response-interval are configurable in the following command contexts:

• router and VPRN service IGMP and MLD command contexts

• router and VPRN service IGMP and MLD group interface command contexts

• IGMP policy and MLD policy command contexts

The IGMP policy and MLD policy query intervals are only applicable to ESM host-based queries. Group interface query intervals are only applicable when the no sub-hosts-only command is configured. This triggers all group interface SAPs to send multicast queries. IGMP policy, MLD policy, and group interface timers use the router IGMP and MLD timers as a default fallback (see ESM host-based queries and Group interface-based queries).

For IGMP, the query-interval must be configured to be longer than the query-last-member-interval and query-response-interval. For MLD, the query-interval must be configured to be longer than the query-last-listener-interval and query-response-interval.

The following ESM multicast replication modes are supported:

• per-host replication

  This mode supports multicast packet replication per subscriber host. Each multicast packet destination uses an individual host source MAC address. The multicast traffic is unicast to the host by replacing the destination MAC from a multicast MAC with the host MAC address. A typical use case for per-host replication is for PPoE subscribers.

• per-SAP replication

  When all subscribers use this replication mode on a single SAP, the system ensures that no more than one multicast packet is sent per (S,G). This mode is used by default when another replication mode is not configured.

  Note: In a BNG CUPS deployment, per-SAP replication is not supported and the system automatically defaults to per-host replication.

• per-SPI replication

  In this mode, the multicast packet replication depends on the number of SLA profile instances configured for the subscriber hosts. A single copy of the multicast packet is sent to hosts that use the same SLA profile, and the multicast packet is replicated for each host that uses a different SLA profile. For example, if all subscriber hosts use the same SLA profile, only one multicast packet is sent from a specific (S,G). However, if the subscriber hosts use three different SLA profiles, three multicast packets are sent.

• multicast redirection

  In this mode, the multicast packets are sent to all the subscribers using a dedicated VLAN.

Use the following CLI syntax to add static group membership entries on an existing SAP (commands for spoke or mesh SDPs are identical):

The following displays an example of a static IGMP snooping configuration on a SAP:

A:ALA-48>config>service>vpls>sap# info
----------------------------------------------
                max-nbr-mac-addr 4
                igmp-snooping
                    fast-leave
                    mrouter-port
                    static
                        group 239.0.10.10
                            source 10.10.10.1
                            source 10.10.10.2
                        exit
                    exit
                exit
----------------------------------------------
A:ALA-48>config>service>vpls>sap#

Use the following CLI syntax to configure Multicast VPLS Registration (MVR). The first step is to register a VPLS as a multicast VPLS.

CLI syntax:

config>service# vpls service-id 
    igmp-snooping
        mvr
            no shutdown
            description description
            group-policy policy-name

Example:

config>service# vpls 1000 
    config>service>vpls# igmp-snooping
    config>service>vpls>snooping# mvr
    config>service>vpls>snooping>mvr# no shutdown
    config>service>vpls>snooping>mvr# description "MVR VPLS"
    config>service>vpls>snooping>mvr# group-policy ‟basic_channels_policy"

The second step is to configure a SAP to take the multicast channels from the registered multicast VPLS.

CLI syntax:

config>service# vpls service-id 
    sap sap-id
        igmp-snooping
            mvr
                from-vpls vpls-id 

Example:

config>service# vpls 1 
    config>service>vpls# sap 1/1/1:100
    config>service>vpls>sap# igmp-snooping
    config>service>vpls>snooping# mvr
    config>service>vpls>snooping>mvr# from-vpls 1000

For MVR by proxy, the destination SAP for the multicast channels should a;sp be configured.

CLI syntax:

config>service# vpls service-id 
    sap sap-id
        igmp-snooping
            mvr
                from-vpls vpls-id 
                to-sap sap-id 

Example:

config>service# vpls 1 
    config>service>vpls# sap 1/1/1:100
    config>service>vpls>sap# igmp-snooping
    config>service>vpls>snooping# mvr
    config>service>vpls>snooping>mvr# from-vpls 1000
    config>service>vpls>snooping>mvr# to-sap 1/1/1:200

See the 7450 ESS, 7750 SR, 7950 XRS, and VSR Multicast Routing Protocols Guide for information about multicast and the commands required to configure basic IGMP and PIM parameters.