2.3.1.1. QoS Policy Propagation Using BGP (QPPB)

This section discusses QPPB as it applies to VPRN, IES, and router interfaces. Refer to the Internet Enhanced Service section and the IP Router Configuration section in the 7750 SR OS Configuration Guide.

QoS policy propagation using BGP (QPPB) is a feature that allows a route to be installed in the routing table with a forwarding-class and priority so that packets matching the route can receive the associated QoS. The forwarding-class and priority associated with a BGP route are set using BGP import route policies. In the industry this feature is called QPPB, and even though the feature name refers to BGP specifically. On SR OS, QPPB is supported for BGP (IPv4, IPv6, VPN-IPv4, VPN-IPv6), RIP and static routes.

While SAP ingress and network QoS policies can achieve the same end result as QPPB, assigning a packet arriving on a particular IP interface to a specific forwarding-class and priority/profile based on the source IP address or destination IP address of the packet the effort involved in creating the QoS policies, keeping them up-to-date, and applying them across many nodes is much greater than with QPPB. In a typical application of QPPB, a BGP route is advertised with a BGP community attribute that conveys a particular QoS. Routers that receive the advertisement accept the route into their routing table and set the forwarding-class and priority of the route from the community attribute.

2.3.1.1.3. Traffic Differentiation Based on Route Characteristics

There may be times when a network operator wants to provide differentiated service to certain traffic flows within its network, and these traffic flows can be identified with known routes. For example, the operator of an ISP network may want to give priority to traffic originating in a particular ASN (the ASN of a content provider offering over-the-top services to the ISP’s customers), following a certain AS_PATH, or destined for a particular next-hop (remaining on-net vs. off-net).

Figure 2 shows an example of an ISP that has an agreement with the content provider managing AS300 to provide traffic sourced and terminating within AS300 with differentiated service appropriate to the content being transported. In this example we presume that ASBR1 and ASBR2 mark the DSCP of packets terminating and sourced, respectively, in AS300 so that other nodes within the ISP’s network do not need to rely on QPPB to determine the correct forwarding-class to use for the traffic. The DSCP or other COS markings could be left unchanged in the ISP’s network and QPPB used on every node.

Figure 2:  Use of QPPB to Differentiate Traffic in an ISP Network 

2.3.1.2. QPPB

There are two main aspects of the QPPB feature:

2.3.1.2.3. Enabling QPPB on an IP Interface

To enable QoS classification of ingress IP packets on an interface based on the QoS information associated with the routes that best match the packets the qos-route-lookup command is necessary in the configuration of the IP interface. The qos-route-lookup command has parameters to indicate whether the QoS result is based on lookup of the source or destination IP address in every packet. There are separate qos-route-lookup commands for the IPv4 and IPv6 packets on an interface, which allows QPPB to enabled for IPv4 only, IPv6 only, or both IPv4 and IPv6. The current QPPB based on a source IP address is not supported for IPv6 packets nor is it supported for ingress subscriber management traffic on a group interface.

The qos-route-lookup command is supported on the following types of IP interfaces:

1. base router network interfaces (config>router>interface)
2. VPRN SAP and spoke SDP interfaces (config>service>vprn>interface)
3. VPRN group-interfaces (config>service>vprn>sub-if>grp-if)
4. IES SAP and spoke SDP interfaces (config>service>ies>interface)
5. IES group-interfaces (config>service>ies>sub-if>grp-if)

When the qos-route-lookup command with the destination parameter is applied to an IP interface and the destination address of an incoming IP packet matches a route with QoS information the packet is classified to the fc and priority associated with that route, overriding the fc and priority/profile determined from the sap-ingress or network qos policy associated with the IP interface. If the destination address of the incoming packet matches a route with no QoS information the fc and priority of the packet remain as determined by the sap-ingress or network qos policy.

Similarly, when the qos-route-lookup command with the source parameter is applied to an IP interface and the source address of an incoming IP packet matches a route with QoS information the packet is classified to the fc and priority associated with that route, overriding the fc and priority/profile determined from the sap-ingress or network qos policy associated with the IP interface. If the source address of the incoming packet matches a route with no QoS information the fc and priority of the packet remain as determined by the sap-ingress or network qos policy.

Currently, QPPB is not supported for ingress MPLS traffic on network interfaces or on CsC PE’-CE’ interfaces (config>service>vprn>nw-if).

Note: QPPB based on a source IP address is not supported for ingress subscriber management traffic on a group interface.

2.3.1.3. QPPB and GRT Lookup

Source-address based QPPB is not supported on any SAP or spoke SDP interface of a VPRN configured with the grt-lookup command.

2.3.1.4. Object Grouping and State Monitoring

This feature introduces a generic operational group object which associates different service endpoints (pseudowires and SAPs) located in the same or in different service instances. The operational group status is derived from the status of the individual components using certain rules specific to the application using the concept. A number of other service entities, the monitoring objects, can be configured to monitor the operational group status and to perform certain actions as a result of status transitions. For example, if the operational group goes down, the monitoring objects will be brought down.

2.3.2.1. IPv6 Enhanced Subscriber Management (ESM)

All IPv6 ESM services require either Routed CO (IES), or Routed CO for VPRN as a supporting service construct. Because of the complexities of the IPv6 link-model, there is currently no support for IPv6 ESM in a VPLS. There is also currently no support for IPv6 in combination with Basic Subscriber Management (BSM). This feature applies to the 7450 ESS and 7750 SR only.

2.3.2.2. RADIUS Accounting

In the 7750 SR OS, the accounting paradigm is based on sla-profile instances, yet this is at odds with traditional RADIUS authentication and accounting which is host-centric. In previous OS releases, it was possible to have many hosts sharing a common sla-profile instance, and thus accounting and QoS parameters. Complications would arise with RADIUS accounting because Accounting-Start and Accounting-Stop are a function of sla-profile instance and not the hosts – this meant that some host-specific parameters (like Framed-Ip-Address) would not be consistently included in RADIUS accounting.

Dual-stack subscribers are now two different hosts sharing a single sla-profile instance. A new RADIUS accounting mode has been introduced to support multiple-host environments.

A new command, host-accounting, is introduced under accounting-policy, which allows configurable behavior.

No host-accounting:

When no host-accounting is configured, accounting behavior is as follows:

Host-accounting enabled:

When host-accounting is configured, additional RADIUS accounting messages are created for host activity in addition to messages for common queue accounting. The behavior is as follows:

This behavior means certain AVP may be in either host, sla-profile instance, or both accounting records. See Table 3.

Note: Interim-Acct records are not sent for hosts, only the start- and stop-accounting messages.

2.3.3.1. Encapsulations

The following SAP encapsulations are supported on the IES services:

2.3.3.2. ATM SAP Encapsulations for IES

The 7750 SR series supports ATM PVC service encapsulation for IES SAPs. Both UNI and NNI cell formats are supported. The format is configurable on a SONET/SDH path basis. A path maps to an ATM VC. All VCs on a path must use the same cell format.

The following ATM encapsulation and transport modes are supported:

2.3.3.3. Pseudowire SAPs

This feature allows customers of an IES, VPRN, or Epipe VLL service and connected to an Ethernet SAP on an Access PE to be backhauled through an Ethernet aggregation network using MPLS pseudowires terminating directly on a Converged PE hosting the IES, VPRN, or Epipe VLL service. If Enhanced Subscriber Management over PW is also used, then the converged PE may also act as a BNG. This service is different from VLL Spoke-SDP termination on an IES or VPRN because access QoS policies can be applied directly at a centralized PE hosting the IES or VPRN instance. This feature uses the same concepts of pseudowire ports and pseudowire SAPs that are used for ESM over MPLS pseudowires, described in the SR OS Triple Play Service Delivery Architecture user guide.

The MPLS pseudowire originates from the first hop aggregation PE (referred to as access PE) upstream of the Access-Node (or directly from a multi-service AN), and terminates on the Converged PE. Multiple customers from a given access-port on the Access-PE can be backhauled over a single MPLS pseudowire towards the Converged PE. This capability allows the network to scale and does not require an MPLS pseudowire per customer between the Access-PE and the Converged PE. The access-port on the Access-PE can be dot1q, q-in-q or NULL encapsulated. The Converged PE terminates the MPLS pseudowire, decapsulates the received frames, and provides access QoS functions including HQoS, without requiring an internal or external loopback. Each MPLS pseudowire is represented on the BNG as a “PW-port” for which SAPs are created. These SAPs are termed “PW SAPs”, and must be statically configured on IES or VPRN interfaces (unlike the ESM case where a capture SAP can be configured). The underlying Ethernet port must be in hybrid mode. Pseudowire SAPs are supported on Ethernet MDAs and on the HSMDAv2.

Figure 3 illustrates the architecture of an aggregation network that uses pseudowire SAPs.

Figure 3:  Network Architecture using Pseudowire SAPs 

2.3.3.4. Encapsulation

The packet is encapsulated on an Ethernet pseudowire, which is associated with a pseudowire port on the Converged PE, and a Spoke-SDP on the access PE. The optional control word is not supported. The SDP could use an LDP LSP, RSVP LSP, BGP RFC3107 tunnel, or LDP over RSVP tunnel. Hash labels are not supported. The SDP may be bound to a port or a LAG, although shaping vports for pseudowire ports on LAGs in distributed mode is not supported. If an SDP is rerouted, then the corresponding pseudowire ports are brought operationally down. Pseudowire ports are associated with an SDP by configuration.

2.3.3.5. Pseudowire SAP Configuration

The following steps are required at the access PE:

The following steps are used to configure a pseudowire SAP on the IES or VPRN service at the Layer 3 PE:

The PW SAP may be mated to an Ethernet SAP or an Ethernet Spoke-SDP in the Epipe VLL service

2.3.3.6. QoS for Pseudowire Ports and Pseudowire SAPs

Pseudowire SAPs support the QoS models allowed for regular VLL, IES or VPRN SAPs. These include:

2.3.3.7. Shaping and Bandwidth Control

Pseudowire SAPs can be shaped on egress by a vport on a physical port. The pseudowire SAP egress cannot explicitly declare which vport to use, but they will inherit the vport used by the pw-port egress shaping.

The vport is represented by a secondary shaper on an HSMDAv2. The intermediate destination identifier, used for ESM on MPLS pseudowires, is not applicable to VLL, IES and VPRN pseudowire SAPs.

If a pseudowire port is configured on a LAG, then vport shaping is only supported if the LAG is in link mode.

Per-access node shaping is configured as follows:

As in the ESM on pseudowire case, vport scheduling on the HSMDAv2 is implemented using an exp-secondary-shaper. This is referred to as a pw-sap-secondary-shaper in the new CLI below. If an hsmda-queue-override secondary-shape is defined for the pw-sap, then the system will use the override, else:

For bandwidth control per pseudowire, the following configuration steps are used:

To make use of the agg-rate rate or port-scheduler under a VPORT, PW SAP queues and schedulers must be configured with the port-parent command. To make use of a scheduler under a VPORT, PW SAP schedulers must be configured with a parent command and the parent-location vport under the tier 1 of the scheduler policy. The egress hierarchical parenting relationship options are shown in Figure 4. See the SR OS Quality of Service guide for more details.

Figure 4:  PW SAP Egress Scheduling Hierarchy Options 

2.3.3.8. Lag Considerations

PW Ports may be bound to VPORT Schedulers bound to a LAG. However, if the LAG is configured in distributed mode, then bandwidth is shared according to the active LAG members across a single IOM. If the LAG spans multiple IOMs, then it effectively operates in link mode across the IOMs. That is, the full LAG bandwidth is allocated to the LAG members on each IOM. Therefore the use of a vport on a distributed mode LAG with a port scheduler on the port or vport and PW SAPs is explicitly not supported and is not a recommended configuration. It is recommended that port-fair mode is used instead.

2.3.3.9. Last Mile Packet Size Adjustment

In the application where pseudowire SAPs are used to apply access QoS for services aggregated from an Ethernet access network, MPLS labels may not be present on the last-mile and link from an access node. In these cases, policers, queues and H-QoS schedulers should account for packets without MPLS overhead, modeled as “encaps-offset”. Vport and port schedulers behave as per the table below. In the data-path, the actual pseudowire encap overhead (taking into account the MPLS labels) added to the packet is tracked, and may be applied to the scheduler calculations via the configured packet-byte-offset.

The exp-secondary-shaper used on the HSMDAv2 always assumes MPLS overhead and does not account for the packet-byte-offset. In all other cases, the rate limit configured for the pseudowire SAP accounts for subscriber or service frame wire rate: without MPLS overhead and including the last mile overhead (unless a packet-byte-offset is configured).

Table 4 summarizes the default packet sizes used at each of the schedulers on the IOM/Ethernet MDA and HSMDAv2, assuming a 1000byte customer packet.

2.3.3.10. Redundancy with Pseudowire SAPs

Within a chassis, IOM and port based redundancy is based on active/backup LAG. The topology for the base MPLS LSP used by the SDP could be constrained such that it could get re-routed in the aggregation network, but would always appear on the LAG ports on the Layer 3 PE. In the case that the tunnel is re-routed to a different port, the MPLS pseudowire SAPs would be brought down.

In order to provide Layer 3 PE redundancy, dual homing of the access PE into separate Layer 3 PEs using active/standby pseudowire status is supported. This is shown in Figure 5.

Figure 5:  Dual Homing into Multiple Layer 3 PEs 

Dual homing operates in a similar manner to Spoke-SDP termination on IES/VPRN. Figure 5 displays the access PE is dual-homed to the Layer 3 PEs using two spoke-SDPs. The endpoint in the access PE is configured to be the master from a pseudowire redundancy perspective using the standby-signaling-master command. The access PE picks one of the Spoke-SDPs to make active, and one to make standby, based on the local configuration of primary or Spoke SDP precedence.

The pseudowire port at the Layer 3 PE behaves as a slave from the perspective of pseudowire status signaling. That is, if its peer signals “PW FWD standby (0x20)” status bit for the given Spoke-SDP and the local configuration does not allow this bit to be ignored, the PE will take the pseudowire port to a local operationally down state. This is consistent with the Spoke-SDP behavior for the case of Spoke-SDP termination on IES/VPRN.

As a consequence, all of the pseudowire SAPs bound to the pseudowire port are taken down, which causes the corresponding IES or VPRN interface to go to a local operationally down state and thus will stop forwarding packets towards this pseudowire port.

Conversely, the formerly standby pseudowire is made active and then the corresponding pseudowire port on the second Layer 3 PE is taken locally operationally up. Therefore, all of the pseudowire SAPs bound to the pseudowire port are brought up, which causes the corresponding IES or VPRN interface to go to a local operationally up state allowing forwarding of packets towards this pseudowire port.

For VLLs, a PW Port always behaves as a slave from the perspective of PW redundancy. This is because the PW Port is taken locally operationally down if any non-zero PW status (including a PW Preferential Forwarding status of 'standby') is received. Support for existing master-slave PW redundancy mechanisms for dual homing of the access PE into separate converged PEs using active/standby PW status is required as shown in Figure 6.

Figure 6:  Master-Slave PW Redundancy  

As in the existing implementation, standby-signaling-master is configured on the Spoke-SDP at the access PE. However explicit configuration of standby-signaling-slave on the PW Port is not required, as this is the default behavior.

The forwarding behavior is the same as when standby-signaling-slave for Epipe Spoke-SDPs. That is, when enabled, if a PW Forwarding Standby (0x20) LDP status message is received for the PW, then the transmit direction is blocked for the PW Port. All PW SAPs bound to the corresponding PW Port are treated from a SAP OAM perspective in the same manner as a fault on the service e.g. an SDP-binding down or remote SAP down.

PW redundancy with multiple active/standby PW Ports or PW SAPs bound to the same Ethernet SAP in the Converged PE is not supported. The Independent Mode of operation for PW Redundancy is not also supported for a PW Port.

2.3.3.11. Operational Group Support for PW Ports

A PW Port state may be linked to the state of an oper-group, such that if the oper-group goes down, the SDP binding for the PW Port will also go operationally down, and thus the corresponding PW status bit signaled (0x00000001 - Pseudowire Not Forwarding). If a status of 0x00000001 is signaled for a currently active PW, and active/standby dual homing is in use then the access PE will fail over to the standby PW to the standby Converged PE.

This is achieved by linking an SDP binding to an operational group for PW SAPs belonging to any supported service types (including those with group interfaces) bound to that PW Port i.e. IES, VPRN, or Epipe VLL. The association to an operational group is configured under the PW port config at the SDP binding level, as follows:

The monitor-oper-group command specifies the operational group to be monitored by the PW-Port under which it is configured. The oper-group name must be already configured under the config>service context before its name is referenced in this command.

The following illustrates how a PW Port can track the status of VPRN uplinks using monitor-oper-group.

Uplinks in a VPRN may be monitored using a BFD session on the network facing IP interfaces in a VPRN or on the network IP interfaces supporting the uplinks.

Oper-groups monitor the state of these BFD sessions inside the VPRN as follows:

Alternatively, the state of network interfaces can be monitored as follows:

The PW Port is then configured with monitor-oper-group as follows:

2.3.4.1. CPE Connectivity Check

Static routes are used within many IES services. Unlike dynamic routing protocols, there is no way to change the state of routes based on availability information for the associated CPE. CPE connectivity check adds flexibility so that unavailable destinations will be removed from the service provider’s routing tables dynamically and minimize wasted bandwidth.

The availability of the far-end static route is monitored through periodic polling. The polling period is configured. If the poll fails a specified number of sequential polls, the static route is marked as inactive.

An ICPM ping mechanism is used to test the connectivity.

If the connectivity check fails and the static route is deactivated, the router will continue to send polls and re-activate any routes that are restored.

2.3.9.1. SRRP Messaging

SRRP uses the same messaging format as VRRP with slight modifications. The source IP address is derived from the system IP address assigned to the local router. The destination IP address and IP protocol are the same as VRRP (224.0.0.18 and 112, respectively).

The message type field is set to 1 (advertisement) and the protocol version is set to 8 to differentiate SRRP message processing from VRRP message processing.

The vr-id field supports an SRRP instance ID of 32 bits.

Due to the large number of subnets backed up by SRRP, only one message every minute carries the gateway IP addresses associated with the SRRP instance. These gateway addresses are stored by the local SRRP instance and are compared with the gateway addresses associated with the local subscriber IP interface.

Unlike VRRP, only two nodes may participate in an SRRP instance due the explicit association between the SRRP instance group IP interface, the associated redundant IP interface and the multi-chassis synchronization (MCS) peering. Since only two nodes are participating, the VRRP skew timer is not utilized when waiting to enter the master state. Also, SRRP always preempts when the local priority is better than the current master and the backup SRRP instance always inherits the master’s advertisement interval from the SRRP advertisement messaging.

SRRP advertisement messages carry a becoming-master indicator flag. The becoming-master flag is set by a node that is attempting to usurp the master state from an existing SRRP master router. When receiving an SRRP advertisement message with a better priority and with the becoming-master flag set, the local master initiates its becoming-backup state, stops routing with the SRRP gateway MAC and sends an SRRP advertisement message with a priority set to zero. The new master continues to send SRRP advertisement messages with the becoming-master flag set until it either receives a return priority zero SRRP advertisement message from the previous master or its becoming-master state timer expires. The new backup node continues to send zero priority SRRP advertisement messages every time it receives an SRRP advertisement message with the becoming-master flag set. After the new master either receives the old masters priority zero SRRP advertisement message or the become-master state timer expires, it enters the master state. The become-master state timer is set to 10 seconds upon entering the become-master state.

The SRRP advertisement message is always evaluated to see if it has a higher priority than the SRRP advertisement that would be sent by the local node. If the advertised priority is equal to the current local priority, the source IP address of the received SRRP advertisement is used as a tie breaker. The node with the lowest IP address is considered to have the highest priority. SRRP will not preempt when priorities are equal. Preemption occurs only when priorities are specified. The lower IP address is only used as a tie-breaker when there is no master in the network. In other words, when both routers are changing from the “init” state, the lower IP will be used to choose the master. If a master already exists, despite having the lower IP address, the system will not preempt the current master.

The SRRP instance maintains the source IP address of the current master. If an advertisement is received with the current masters source IP address and the local priority is higher priority than the masters advertised priority, the local node immediately enters the becoming-master state unless the advertised priority is zero. If the advertised priority is zero, the local node bypasses the becoming-master state and immediately enters the master state. Priority zero is a special case and is sent when an SRRP instance is relinquishing the master state.

2.3.9.2. SRRP and Multi-Chassis Synchronization

In order to take full advantage of SRRP resiliency and diagnostic capabilities, the SRRP instance should be tied to a MCS peering that terminates on the redundant node. The SRRP instance is tied to the peering using the srrp srrp-id command within the appropriate MCS peering configuration. Once the peering is associated with the SRRP instance, MCS will synchronize the local information about the SRRP instance with the neighbor router. MCS automatically derives the MCS key for the SRRP instance based on the SRRP instance ID. For example, an SRRP instance ID of 1 would appear in the MCS peering database with a MCS-key srrp-0000000001.

The SRRP instance information stored and sent to the neighbor router consists of:

2.3.9.3. SRRP Instance

The SRRP instance uses the received information to verify provisioning and obtain operational status of the SRRP instance on the neighboring router.

2.3.9.4. Subscriber Subnet Owned IP Address Connectivity

In order for the network to reliably reach the owned IP addresses on a subscriber subnet, it is not necessary for the owning node to advertise the IP addresses as /32 host routes into the core. Network reachability to the subscriber subnet is advertised into the IGP core by both of the dual homing nodes. The shortest path to the subscriber may not always traverse the active path for a subscriber. In this case, the path traverses the non-active/primary node for the subscriber and the traffic will be redirected through the redundant interface to the other node through the redundant interface to the active path. This ensures that all downstream traffic to a given subscriber will always flow through one node.

2.3.9.5. Subscriber Subnet SRRP Gateway IP Address Connectivity

The SRRP gateway IP addresses on the subscriber subnets cannot be advertised as /32 host routes since they may be active (master) on multiple group IP interfaces on multiple SRRP routers. Without a /32 host route path, the network will forward any packet destined to an SRRP gateway IP address to the closest router advertising the subscriber subnet. While a case may be made that only a node that is currently forwarding for the gateway IP address in a master state should respond to ping or other diagnostic messages, the distribution of the subnet and the case of multiple masters make any resulting response or non-response inconclusive at best. To provide some ability to ping the SRRP gateway address from the network side reliably, any node receiving the ICMP ping request responds if the gateway IP address is defined on its subscriber subnet.

2.3.9.6. Receive SRRP Advertisement SAP and Anti-Spoof

The group IP interface SAPs are designed to support subscriber hosts and perform an ingress anti-spoof function that ensures that any IP packet received on the group IP interface is coming in the correct SAP with the correct MAC address. If the IP and MAC are not registered as valid subscriber hosts on the SAP, the packet is silently discarded. Since the SRRP advertisement source IP addresses are not subscriber hosts, an anti-spoof entry will not exist and SRRP advertisement messages would normally be silently discarded. To avoid this issue, when a group IP interface SAP is configured to send and receive SRRP advertisement messages, anti-spoof processing on the SAP is disabled. This precludes subscriber host management on the SRRP messaging SAP.

2.3.9.7. BFD with SRRP/VRRP

BFD with SRRP is supported. This allows the use of longer timers inside SRRP resulting in more SRRP instances while still retaining fast failure detection with BFD.