MPLS-TE enables ISPs to route network traffic to offer the best service to their users in terms of throughput and delay. By making the service provider more efficient, traffic engineering reduces the cost of the network.

Currently, some ISPs base their services on an overlay model. In the overlay model, transmission facilities are managed by Layer 2 switching. The routers see only a fully meshed virtual topology, making most destinations appear one hop away. If you use the explicit Layer 2 transit layer, you can precisely control how traffic uses available bandwidth. However, the overlay model has numerous disadvantages. MPLS-TE achieves the TE benefits of the overlay model without running a separate network and without a non-scalable, full mesh of router interconnects.

MPLS-TE automatically establishes and maintains label switched paths (LSPs) across the backbone by using RSVP. The path that an LSP uses is determined by the LSP resource requirements and network resources, such as bandwidth. Available resources are flooded by means of extensions to a link-state-based Interior Gateway Protocol (IGP).

MPLS-TE tunnels are calculated at the LSP headend router, based on a fit between the required and available resources (constraint-based routing). The IGP automatically routes the traffic to these LSPs.

One approach to engineering a backbone is to define a mesh of tunnels from every ingress device to every egress device. The MPLS-TE path calculation and signaling modules determine the path taken by the LSPs for these tunnels, subject to resource availability and the dynamic state of the network.

The IGP (operating at an ingress device) determines which traffic should go to which egress device, and steers that traffic into the tunnel from ingress to egress. A flow from an ingress device to an egress device might be so large that it cannot fit over a single link, so it cannot be carried by a single tunnel. In this case, multiple tunnels between a given ingress and egress can be configured, and the flow is distributed using load sharing among the tunnels.

Note	GRE over MPLS-TE tunnel is not supported. Hence, you cannot carry GRE traffic over an LSP established for MPLS-TE tunnel using RSVP-TE. This restriction also applies to SR-TE tunnels.

The MPLS Traffic Engineering AutoTunnel Backup feature enables a router to dynamically build backup tunnels on the interfaces that are configured with MPLS TE tunnels. This feature enables a router to dynamically build backup tunnels when they are needed. This prevents you from having to build MPLS TE tunnels statically .

The MPLS Traffic Engineering (TE)—AutoTunnel Backup feature has these benefits:

• Backup tunnels are built automatically, eliminating the need for users to preconfigure each backup tunnel and then assign the backup tunnel to the protected interface.

• Protection is expanded—FRR does not protect IP traffic that is not using the TE tunnel or Label Distribution Protocol (LDP) labels that are not using the TE tunnel.

This feature protects against these failures:

• P2P Tunnel NHOP protection —Protects against link failure for the associated P2P protected tunnel

• P2P Tunnel NNHOP protection —Protects against node failure for the associated P2P protected tunnel

• P2MP Tunnel NHOP protection —Protects against link failure for the associated P2MP protected tunnel

The periodic backup promotion attempts to find and assign a better backup for primary tunnels that are already protected.

With AutoTunnel Backup, the only scenario where two backups can protect the same primary tunnel is when both an NHOP and NNHOP AutoTunnel Backups get created. The backup assignment takes place as soon as the NHOP and NNHOP backup tunnels come up. So, there is no need to wait for the periodic promotion.

Although there is no exception for AutoTunnel Backups, periodic backup promotion has no impact on primary tunnels protected by AutoTunnel Backup.

One exception is when a manual promotion is triggered by the user using the mpls traffic-eng fast-reroute timers promotion command, where backup assignment or promotion is triggered on all FRR protected primary tunnels--even unprotected ones. This may trigger the immediate creation of some AutoTunnel Backup, if the command is entered within the time window when a required AutoTunnel Backup has not been yet created.

You can configure the periodic promotion timer using the global configuration mpls traffic-eng fast-reroute timers promotion sec command. The range is 0 to 604800 seconds.

Note	A value of 0 for the periodic promotion timer disables the periodic promotion.

Prestandard DS-TE uses the Cisco proprietary mechanisms for RSVP signaling and IGP advertisements. This DS-TE mode does not interoperate with third-party vendor equipment. Note that prestandard DS-TE is enabled only after configuring the sub-pool bandwidth values on MPLS-enabled interfaces.

Prestandard Diff-Serve TE mode supports a single bandwidth constraint model a Russian Doll Model (RDM) with two bandwidth pools: global-pool and sub-pool.

TE class map is not used with Prestandard DS-TE mode.

IETF DS-TE mode uses IETF-defined extensions for RSVP and IGP. This mode interoperates with third-party vendor equipment.

IETF mode supports multiple bandwidth constraint models, including RDM and MAM, both with two bandwidth pools. In an IETF DS-TE network, identical bandwidth constraint models must be configured on all nodes.

TE class map is used with IETF DS-TE mode and must be configured the same way on all nodes in the network.

IETF DS-TE mode provides support for the RDM and MAM bandwidth constraints models. Both models support up to two bandwidth pools.

Cisco IOS XR software provides global configuration for the switching between bandwidth constraint models. Both models can be configured on a single interface to preconfigure the bandwidth constraints before swapping to an alternate bandwidth constraint model.

Note	NSF is not guaranteed when you change the bandwidth constraint model or configuration information.

By default, RDM is the default bandwidth constraint model used in both pre-standard and IETF mode.

Each of the eight available bandwidth values advertised in the IGP corresponds to a TE class. Because the IGP advertises only eight bandwidth values, there can be a maximum of only eight TE classes supported in an IETF DS-TE network.

TE class mapping must be exactly the same on all routers in a DS-TE domain. It is the responsibility of the operator configure these settings properly as there is no way to automatically check or enforce consistency.

The operator must configure TE tunnel class types and priority levels to form a valid TE class. When the TE class map configuration is changed, tunnels already up are brought down. Tunnels in the down state, can be set up if a valid TE class map is found.

The default TE class and attributes are listed. The default mapping includes four class types.

TE Link Management (TE-Link) notifies IGP for both global pool and sub-pool available bandwidth and maximum bandwidth to flood the network in these events:

• Periodic timer expires (this does not depend on bandwidth pool type).

• Tunnel origination node has out-of-date information for either available global pool or sub-pool bandwidth, causing tunnel admission failure at the midpoint.

• Consumed bandwidth crosses user-configured thresholds. The same threshold is used for both global pool and sub-pool. If one bandwidth crosses the threshold, both bandwidths are flooded.

Flooding frequently can burden a network because all routers must send out and process these updates. Infrequent flooding causes tunnel heads (tunnel-originating nodes) to have out-of-date information, causing tunnel admission to fail at the midpoints.

You can control the frequency of flooding by configuring a set of thresholds. When locked bandwidth (at one or more priority levels) crosses one of these thresholds, flooding is triggered.

Thresholds apply to a percentage of the maximum available bandwidth (the global pool), which is locked, and the percentage of maximum available guaranteed bandwidth (the sub-pool), which is locked. If, for one or more priority levels, either of these percentages crosses a threshold, flooding is triggered.

Note

Setting up a global pool TE tunnel can cause the locked bandwidth allocated to sub-pool tunnels to be reduced (and hence to cross a threshold). A sub-pool TE tunnel setup can similarly cause the locked bandwidth for global pool TE tunnels to cross a threshold. Thus, sub-pool TE and global pool TE tunnels can affect each other when flooding is triggered by thresholds.

The MPLS-TE interarea tunneling feature allows you to establish P2P and P2MP TE tunnels spanning multiple Interior Gateway Protocol (IGP) areas and levels, thereby eliminating the requirement that headend and tailend routers reside in a single area.

Interarea support allows the configuration of a TE LSP that spans multiple areas, where its headend and tailend label switched routers (LSRs) reside in different IGP areas.

Multiarea and Interarea TE are required by the customers running multiple IGP area backbones (primarily for scalability reasons). This lets you limit the amount of flooded information, reduces the SPF duration, and lessens the impact of a link or node failure within an area, particularly with large WAN backbones split in multiple areas.

Multiarea support allows an area border router (ABR) LSR to support MPLS-TE in more than one IGP area. A TE LSP is still confined to a single area.

Multiarea and Interarea TE are required when you run multiple IGP area backbones. The Multiarea and Interarea TE allows you to:

• Limit the volume of flooded information.

• Reduce the SPF duration.

• Decrease the impact of a link or node failure within an area.

As shown in the figure, R2, R3, R7, and R4 maintain two databases for routing and TE information. For example, R3 has TE topology information related to R2, flooded through Level-1 IS-IS LSPs plus the TE topology information related to R4, R9, and R7, flooded as Level 2 IS-IS Link State PDUs (LSPs) (plus, its own IS-IS LSP).

Note	You can configure multiple areas within an IS-IS Level 1. This is transparent to TE. TE has topology information about the IS-IS level, but not the area ID.

Loose hop optimization allows the reoptimization of tunnels spanning multiple areas and solves the problem which occurs when an MPLS-TE LSP traverses hops that are not in the LSP's headend's OSPF area and IS-IS level.

Interarea MPLS-TE allows you to configure an interarea traffic engineering (TE) label switched path (LSP) by specifying a loose source route of ABRs along the path. It is the then the responsibility of the ABR (having a complete view of both areas) to find a path obeying the TE LSP constraints within the next area to reach the next hop ABR (as specified on the headend). The same operation is performed by the last ABR connected to the tailend area to reach the tailend LSR.

For P2MP-TE tunnels, ABRs support loose hop ERO expansion to find path to the next ABR until it reaches to the tail-end LSR, without introducing remerge.

You must be aware of these considerations when using loose hop optimization:

• You must specify the router ID of the ABR node (as opposed to a link address on the ABR).

• When multiarea is deployed in a network that contains subareas, you must enable MPLS-TE in the subarea for TE to find a path when loose hop is specified.

• You must specify the reachable explicit path for the interarea tunnel.

Loose hop reoptimization allows the reoptimization of the tunnels spanning multiple areas and solves the problem which occurs when an MPLS-TE headend does not have visibility into other IGP areas.

Whenever the headend attempts to reoptimize a tunnel, it tries to find a better path to the ABR in the headend area. If a better path is found then the headend initiates the setup of a new LSP. In case a suitable path is not found in the headend area, the headend initiates a querying message. The purpose of this message is to query the ABRs in the areas other than the headend area to check if there exist any better paths in those areas. The purpose of this message is to query the ABRs in the areas other than the headend area, to check if a better path exists. If a better path does not exist, ABR forwards the query to the next router downstream. Alternatively, if better path is found, ABR responds with a special Path Error to the headend to indicate the existence of a better path outside the headend area. Upon receiving the Path Error that indicates the existence of a better path, the headend router initiates the reoptimization.

Because one IGP area does not have visibility into another IGP area, it is not possible to assign backup to protect ABR node. To overcome this problem, node ID sub-object is added into the record route object of the primary tunnel so that at a PLR node, backup destination address can be checked against primary tunnel record-route object and assign a backup tunnel.

If a link failure occurs within an area, the upstream router directly connected to the failed link generates an RSVP path error message to the headend. As a response to the message, the headend sends an RSVP path tear message and the corresponding path option is marked as invalid for a specified period and the next path-option (if any) is evaluated.

To retry the ABR immediately, a second path option (identical to the first one) should be configured. Alternatively, the retry period (path-option hold-down, 2 minutes by default) can be tuned to achieve a faster retry.

TE tunnel interfaces are advertised in the IGP network just like any other links. Routers can then use these advertisements in their IGPs to compute the SPF even if they are not the head end of any TE tunnels.

The MPLS-TE Forwarding Adjacency feature has these restrictions:

• Using the MPLS-TE Forwarding Adjacency increases the size of the IGP database by advertising a TE tunnel as a link.

• The MPLS-TE Forwarding Adjacency is supported by Intermediate System-to-Intermediate System (IS-IS).

• When the MPLS-TE Forwarding Adjacency is enabled on a TE tunnel, the link is advertised in the IGP network as a Type-Length-Value (TLV) 22 without any TE sub-TLV.

• MPLS-TE forwarding adjacency tunnels must be configured bidirectionally.

• Multicast intact is not supported with MPLS-TE Forwarding Adjacency.

Your network must support the following features before enabling the MPLS -TE Forwarding Adjacency feature:

• MPLS

• IP Cisco Express Forwarding

• Intermediate System-to-Intermediate System (IS-IS)

• OSPF

Policy-Based Tunnel Selection (PBTS) provides a mechanism that lets you direct traffic into specific TE tunnels based on different criteria. PBTS will benefit Internet service providers (ISPs) who carry voice and data traffic through their MPLS and MPLS/VPN networks, who want to route this traffic to provide optimized voice service.

PBTS works by selecting tunnels based on the classification criteria of the incoming packets, which are based on the IP precedence, experimental (EXP), differentiated services code point (DSCP), or type of service (ToS) field in the packet. Default-class configured for paths is always zero (0). If there is no TE for a given forward-class, then the default-class (0) will be tried. If there is no default-class, then the packet is dropped.PBTS supports up to seven (exp 1 - 7) EXP values associated with a single TE tunnel.

In PBTS, a class-map is defined for various types of packet, and associating this class-map with a forward-class. The class-map defines the matching criteria for classifying a particular type of traffic, while the forward-class defines the forwarding path these packets should take. Once a class-map is associated with a forwarding-class in the policy map, all the packets that match the class-map are forwarded as defined in the policy-map. The egress traffic engineering (TE) tunnel interfaces that the packets should take for each forwarding-class is specified by associating the TE interface explicitly, or implicitly in the case of default value, with the forward-group. When the TE interfaces are associated with the forward-class, they can be exported to the routing protocol module using the auto-route command, which will then associate the route in the FIB database with these tunnels. If the TE interface is not explicitly associated with a forward-class, it gets associated with a default-class (0). All non-TE interfaces through the destination route was learned will be pushed down by routing protocol to the forwarding plane with forwarding class set to default-class.

When PBR is configured, TE tunnel interfaces are selected to forward traffic based on matching class-types in policy-map. When a forward-class value for a TE tunnel is configured, that TE process pass to LSD process as part of the label rewrite for the tunnel head. The TE allows one 32-bit value for a forward-class per tunnel that is opaque to TE and TE will use it to program the tunnel in forwarding.

PBTS supports a maximum of eight forward-class and eight TE tunnels with in each forward-class. A maximum of 32 TE tunnels can be associated with the destination route.

Forward-class configuration is supported for auto-mesh tunnels and forward-class can be configured in the attribute-set used by the auto-mesh tunnels. Changing the forward-class configuration does not affect the tunnel state. TE updates the forwarding with the new forward-class value. For TE path protection, both primary and standby LSPs use the same forward-class value for the tunnel. Unequal load-sharing of traffic is supported for the ECMP TE tunnels across forward-classes.

Forward-class configuration does not apply to the FRR backup tunnels and will be ignored. The forward-class configuration is not supported for the auto-backup tunnels and P2MP-TE (MTE) tunnels.

Policy Based Tunnel Selection (PBTS) provides a mechanism that directs traffic into TE tunnels based on incoming packets TOS/EXP bits. The PBTS default class enhancement can be explained as follows:

• Add a new class called default so that you can configure a tunnel of class (1-7 or default). You can configure more than one default tunnels. By default, tunnels of class 0 no longer serves as default tunnel.

• The control plane can pick up to 8 default tunnels to carry default traffic.

• The forwarding plane applies the same load-balancing logic on the default tunnels such that default traffic load is shared over them.

• Default tunnels are not used to forward traffic if each class of traffic is served by at least one tunnel of the respective class.

• A tunnel is implicitly assigned to class 0 if the tunnel is not configured with a specific class.

• If no default tunnel is available for forwarding, the traffic is dropped.

• Both LDP and IGP paths are assigned to a new default class. LDP and IGP no longer statically associate to class 0 in the platforms, which support this new default class enhancement.

This PBTS feature enhancement allows you to configure forward class zero or the default class as a PBTS forwarding class effective with Cisco IOS-XR release 6.4.1. Earlier, only values from one to seven were configurable as PBTS forwarding classes.

The Set DF Bit feature enables to apply 'set df (do not fragment)' policy to an interface. Any packet that matches with the set df policy will either clear the bit or set the bit.

The set df bit policy can be enabled to clear the df bit before forwarding the packet in IPv4 traffic.

For more information on Set DF Bit, see the Implementing MPLS Traffic Engineering chapter in the MPLS Configuration Guide for Cisco ASR 9000 Series Routers. For more information on Set DF Bit commands, see the MPLS Traffic Engineering Commands chapter in the MPLS Command Reference for Cisco ASR 9000 Series Routers.

Service Path Preference (SPP) includes the following features:

• SPP and BGP Flow Spec are mutually exclusive

• SPP is service agnostic and can be applied to L3VPN services on an MPLS or Segment Routing core network

• SPP provides the facility to separate L3VPN services in dedicated LSPs or tunnels within the core transport network based on SLA. Therefore service delivery on PE devices is simplified.

• If you configure SPP in a setup where PBTS is already configured, the PBTS configuration is given preference. If a packet matches the PBTS lookup, the specified forwarding class is used and SPP configuration is ignored.

• SPP extends the concept PBTS by associating a control plane policy with a forwarding class. The route policy is configured to use a forward-class when specific conditions are met

• SPP is supported for interface paths such as Traffic-Engineering Tunnels, RSVP-TE, and Segment Routing Tunnels (SR-TE).

SPP allows services to select a path based on policies configured in the control plane or based on the preferences set in the egress PE node.

Consider a scenario where you have two Provider Edge (PE) routers in a setup. PE1 functions as ingress node and PE2 functions as egress node.

The egress PE (PE2) receives the routes from the customers and assigns the routes with prefixes such as BGP community, next hop attribute or vrf attribute. The local policies determine the attribute to be assigned to the customer.

PE1 associates a forward-class to the prefix based on the local policies that are created based on a combination of vrf, address-family, next-hop, and community, to match a forward-class. The pre-configured tunnel with matching forward-class is selected for forwarding the traffic.

The following samples explain how to configure route policies for service path preference:

MPLS-TE automatic bandwidth is configured on individual Label Switched Paths (LSPs) at every head-end. MPLS-TE monitors the traffic rate on a tunnel interface. Periodically, MPLS-TE resizes the bandwidth on the tunnel interface to align it closely with the traffic in the tunnel. MPLS-TE automatic bandwidth can perform these functions:

• Monitors periodic polling of the tunnel output rate

• Resizes the tunnel bandwidth by adjusting the highest rate observed during a given period

For every traffic-engineered tunnel that is configured for an automatic bandwidth, the average output rate is sampled, based on various configurable parameters. Then, the tunnel bandwidth is readjusted automatically based upon either the largest average output rate that was noticed during a certain interval, or a configured maximum bandwidth value.

This table lists the automatic bandwidth functions.

The output rate on a tunnel is collected at regular intervals that are configured by using the application command in MPLS-TE auto bandwidth interface configuration mode. When the application period timer expires, and when the difference between the measured and the current bandwidth exceeds the adjustment threshold, the tunnel is reoptimized. Then, the bandwidth samples are cleared to record the new largest output rate at the next interval.

When reoptimizing the LSP with the new bandwidth, a new path request is generated. If the new bandwidth is not available, the last good LSP continues to be used. This way, the network experiences no traffic interruptions.

If minimum or maximum bandwidth values are configured for a tunnel, the bandwidth, which the automatic bandwidth signals, stays within these values.

Note	When more than 100 tunnels are auto-bw enabled, the algorithm will jitter the first application of every tunnel by a maximum of 20% (max 1hour). The algorithm does this to avoid too many tunnels running auto bandwidth applications at the same time.

If a tunnel is shut down, and is later brought again, the adjusted bandwidth is lost and the tunnel is brought back with the initial configured bandwidth. In addition, the application period is reset when the tunnel is brought back.

Adjustment Threshold is defined as a percentage of the current tunnel bandwidth and an absolute (minimum) bandwidth. Both thresholds must be fulfilled for the automatic bandwidth to resignal the tunnel. The tunnel bandwidth is resized only if the difference between the largest sample output rate and the current tunnel bandwidth is larger than the adjustment thresholds.

For example, assume that the automatic bandwidth is enabled on a tunnel in which the highest observed bandwidth B is 30 Mbps. Also, assume that the tunnel was initially configured for 45 Mbps. Therefore, the difference is 15 mbit/s. Now, assuming the default adjustment thresholds of 10% and 10kbps, the tunnel is signalled with 30 Mbps when the application timer expires. This is because 10% of 45Mbit/s is 4.5 Mbit/s, which is smaller than 15 Mbit/s. The absolute threshold, which by default is 10kbps, is also crossed.

Overflow detection is used if a bandwidth must be resized as soon as an overflow condition is detected, without having to wait for the expiry of an automatic bandwidth application frequency interval.

For overflow detection one configures a limit N, a percentage threshold Y% and optionally, a minimum bandwidth threshold Z. The percentage threshold is defined as the percentage of the actual signalled tunnel bandwidth. When the difference between the measured bandwidth and the actual bandwidth are both larger than Y% and Z threshold, for N consecutive times, then the system triggers an overflow detection.

The bandwidth adjustment by the overflow detection is triggered only by an increase of traffic volume through the tunnel, and not by a decrease in the traffic volume. When you trigger an overflow detection, the automatic bandwidth application interval is reset.

By default, the overflow detection is disabled and needs to be manually configured.

Underflow detection is used when the bandwidth on a tunnel drops significantly, which is similar to overflow but in reverse.

Underflow detection applies the highest bandwidth value from the samples which triggered the underflow. For example, if you have an underflow limit of three, and the following samples trigger the underflow for 10 kbps, 20 kbps, and 15 kbps, then, 20 kbps is applied.

Unlike overflow, the underflow count is not reset across an application period. For example, with an underflow limit of three, you can have the first two samples taken at the end of an application period and then the underflow gets triggered by the first sample of the next application period.

The Point-to-Multipoint (P2MP) Resource Reservation Protocol-Traffic Engineering (RSVP-TE) solution allows service providers to implement IP multicast applications, such as IPTV and real-time video, broadcast over the MPLS label switch network. The RSVP-TE protocol is extended to signal point-to-point (P2P) and P2MP label switched paths (LSPs) across the MPLS networks.

By using RSVP-TE extensions as defined in RFC 4875, multiple subLSPs are signaled for a given TE source. The P2MP tunnel is considered as a set of Source-to-Leaf (S2L) subLSPs that connect the TE source to multiple leaf Provider Edge (PE) nodes.

At the TE source, the ingress point of the P2MP-TE tunnel, IP multicast traffic is encapsulated with a unique MPLS label, which is associated with the P2MP-TE tunnel. The traffic continues to be label-switched in the P2MP tree. If needed, the labeled packet is replicated at branch nodes along the P2MP tree. When the labeled packet reaches the egress leaf (PE) node, the MPLS label is removed and forwarded onto the IP multicast tree across the PE-CE link.

To enable end-to-end IP multicast connectivity, RSVP is used in the MPLS-core for P2MP-TE signaling and PIM is used for PE-CE link signaling.

These are the definitions for Point-to-Multipoint (P2MP) tunnels:

Source

Configures the node in which Label Switched Path (LSP) signaling is initiated.

Mid-point

Specifies the transit node in which LSP signaling is processed (for example, not a source or receiver).

Receiver, Leaf, and Destination

Specifies the node in which LSP signaling ends.

Branch Point

Specifies the node in which packet replication is performed.

Bud Node

Specifies the node that not only acts as a transit for some S2Ls but also acts as a termination point for a S2L of a P2MP TE tunnel.

Source-to-Leaf (S2L) SubLSP

Specifies the P2MP-TE LSP segment that runs from the source to one leaf.

RSVP-TE signals a P2MP tunnel base that is based on a manual configuration. If all Source-to-Leaf (S2L)s use an explicit path, the P2MP tunnel creates a static tree that follows a predefined path based on a constraint such as a deterministic Label Switched Path (LSP). If the S2L uses a dynamic path, RSVP-TE creates a P2MP tunnel base on the best path in the RSVP-TE topology. RSVP-TE supports bandwidth reservation for constraint-based routing.

When an explicit path option is used, specify both the local and peer IP addresses in the explicit path option, provided the link is a GigabitEthernet or a TenGigE based interface. For point-to-point links like POS or bundle POS, it is sufficient to mention the remote or peer IP address in the explicit path option.

RSVP-TE distributes stream information in which the topology tree does not change often (where the source and receivers are). For example, large scale video distribution between major sites is suitable for a subset of multicast applications. Because multicast traffic is already in the tunnel, the RSVP-TE tree is protected as long as you build a backup path.

Fast-Reroute (FRR) capability is supported for P2MP RSVP-TE by using the unicast link protection. You can choose the type of traffic to go to the backup link.

The P2MP tunnel is applicable for all TE Tunnel destination (IntraArea and InterArea ). Inter-AS is not supported.

The P2MP tunnel is signaled by the dynamic and explicit path option in the IGP intra area. Only interArea and interAS, which are used for the P2MP tunnels, are signaled by the verbatim path option.

MPLS-TE Fast Reroute (FRR) is a mechanism to minimize interruption in traffic delivery to a TE Label Switched Path (LSP) destination as a result of link failures. FRR enables temporarily fast switching of LSP traffic along an alternative backup path around a network failure, until the TE tunnel source signals a new end-to-end LSP.

Both Point-to-Point (P2P) and P2MP-TE support only the Facility FRR method from RFC 4090.

P2P LSPs are used to backup  P2MP S2L ( source 2  Leaf ). Only link and bandwidth protection for P2MP S2Ls are supported. Node protection is not supported.

MPLS-TE link protection relies on the fact that labels for all primary LSPs and subLSPs are using the MPLS global label allocation. For example, one single (global) label space is used for all MPLS-TE enabled physical interfaces on a given MPLS LSP.

The Point-to-Multipoint Label Switch Path (P2MP LSP) has only a single root, which is the Ingress Label Switch Router (LSR). The P2MP LSP is created based on a receiver that is connected to the Egress LSR. The Egress LSR initiates the creation of the tree (for example, tunnel grafting or pruning is done by performing an individual sub-LSP operation) by creating the Forwarding Equivalency Class (FEC) and Opaque Value.

Note	Grafting and pruning operate on a per destination basis.

The Opaque Value contains the stream information that uniquely identifies the tree to the root. To receive label switched multicast packets, the Egress Provider Edge (PE) indicates to the upstream router (the next hop closest to the root) which label it uses for the multicast source by applying the label mapping message.

The upstream router does not need to have any knowledge of the source; it needs only the received FEC to identify the correct P2MP LSP. If the upstream router does not have any FEC state, it creates it and installs the assigned downstream outgoing label into the label forwarding table. If the upstream router is not the root of the tree, it must forward the label mapping message to the next hop upstream. This process is repeated hop-by-hop until the root is reached.

By using downstream allocation, the router that wants to receive the multicast traffic assigns the label for it. The label request, which is sent to the upstream router, is similar to an unsolicited label mapping (that is, the upstream does not request it). The upstream router that receives that label mapping uses the specific label to send multicast packets downstream to the receiver. The advantage is that the router, which allocates the labels, does not get into a situation where it has the same label for two different multicast sources. This is because it manages its own label space allocation locally.

P2MP tunnels are signaled by using the dynamic and explicit path-options in an IGP intra area. InterArea cases for P2MP tunnels are signaled by the verbatim path option.

Path options for P2MP tunnels are individually configured for each sub-LSP. Only one path option per sub-LSP (destination) is allowed. You can choose whether the corresponding sub-LSP is dynamically or explicitly routed. For the explicit option, you can configure the verbatim path option to bypass the topology database lookup and verification for the specified destination.

Both dynamic and explicit path options are supported on a per destination basis by using the path-option (P2MP-TE) command. In addition, you can combine both path options.

Explicit Path Option

Configures the intermediate hops that are traversed by a sub-LSP going from the TE source to the egress MPLS node. Although an explicit path configuration enables granular control sub-LSP paths in an MPLS network, multiple explicit paths are configured for specific network topologies with a limited number of (equal cost) links or paths.

Dynamic Path Option

Computes the IGP path of a P2MP tree sub-LSP that is based on the OSPF and ISIS algorithm. The TE source is dynamically calculated based on the IGP topology.

Note	Dynamic path option can only compute fully-diverse standby paths. While, explicit path option supports partially diverse standby paths as well.

The Explicit Path configuration allows you to configure the explicit path. An IP explicit path is a list of IP addresses, each representing a node or link in the explicit path.

The MPLS Traffic Engineering (TE)—IP Explicit Address Exclusion feature provides a means to exclude a link or node from the path for an Multiprotocol Label Switching (MPLS) TE label-switched path (LSP).

This feature is enabled through the explicit-path command that allows you to create an IP explicit path and enter a configuration submode for specifying the path. The feature adds to the submode commands of the exclude-address command for specifying addresses to exclude from the path.

The feature also adds to the submode commands of the exclude-srlg command that allows you to specify the IP address to get SRLGs to be excluded from the explicit path.

If the excluded address or excluded srlg for an MPLS TE LSP identifies a flooded link, the constraint-based shortest path first (CSPF) routing algorithm does not consider that link when computing paths for the LSP. If the excluded address specifies a flooded MPLS TE router ID, the CSPF routing algorithm does not allow paths for the LSP to traverse the node identified by the router ID.

Fast ReRoute (FRR) protects MPLS TE Label Switch Paths (LSPs) from link and node failures by locally repairing the LSPs at the point of failure. This protection allows data to continue to flow on LSPs, while their headend routers attempt to establish new end-to-end LSPs to replace them. FRR locally repairs the protected LSPs by rerouting them over backup tunnels that bypass failed links or nodes.

Backup tunnels that bypass only a single link of the LSP's path provide Link Protection. They protect LSPs by specifying the protected link IP addresses to extract SRLG values that are to be excluded from the explicit path, thereby bypassing the failed link. These are referred to as next-hop (NHOP) backup tunnels because they terminate at the LSP's next hop beyond the point of failure. Figure 1 illustrates an NHOP backup tunnel.

In the topology shown in the above figure, the backup tunnel path computation can be performed in this manner:

• Get all SRLG values from the exclude-SRLG link (SRLG values 5 and 6)

• Mark all the links with the same SRLG value to be excluded from SPF

• Path computation as CSPF R2->R6->R7->R3

FRR provides Node Protection for LSPs. Backup tunnels that bypass next-hop nodes along LSP paths are called NNHOP backup tunnels because they terminate at the node following the next-hop node of the LSP paths, thereby bypassing the next-hop node. They protect LSPs when a node along their path fails, by enabling the node upstream to the point of failure to reroute the LSPs and their traffic, around the failed node to the next-next hop. They also protect LSPs by specifying the protected link IP addresses that are to be excluded from the explicit path, and the SRLG values associated with the IP addresses excluded from the explicit path. NNHOP backup tunnels also provide protection from link failures by bypassing the failed link as well as the node. Figure 2 illustrates an NNHOP backup tunnel.

In the topology shown in the above figure, the backup tunnel path computation can be performed in this manner:

• Get all SRLG values from the exclude-SRLG link (SRLG values 5 and 6)

• Mark all links with the same SRLG value to be excluded from SPF

• Verify path with SRLG constraint

• Path computation as CSPF R2->R9->R10->R4

This section describes the following:

• Delivery of Packets During a Failure

• Multiple Backup Tunnels Protecting the Same Interface

In shared-risk link groups (SRLG) fate-sharing, links are assigned one or more numbers to represent risks. When two links are assigned a common number then this indicates that these two links are sharing fate. In the weighted-SRLG auto-backup path computation mode, the links that share SRLG numbers with the protected link are not excluded from the topology. The admin-weight of these links is set to reflect the sharing of SRLG with the protected link. Setting the admin weight consists of adding a penalty metric to make using the link less desirable.

For more information about Weighted-SRLG auto-backup path computation, see Implementing MPLS Traffic Engineering chapter in the MPLS Configuration Guide for Cisco ASR 9000 Series Routers. For more information about Weighted-SRLG auto-backup path computation, see MPLS Traffic Engineering Commands chapter in the MPLS Command Reference for Cisco ASR 9000 Series Routers.

There are few limitations to the configured SRLG feature:

• The exclude-address and exclude-srlg options are not allowed in the IP explicit path strict-address network.

• Whenever SRLG values are modified after tunnels are signalled, they are verified dynamically in the next path verification cycle.

MPLS Traffic Engineering Shared Risk Link Groups (SRLG) feature has been enhanced to support:

• Increase from 32 to 64 (59 for ISIS) groups.

• Increase from 250 to 500 interfaces.

You can specify a value for an attribute within a path option attribute-set template. This does not prevent the configuring of the same attribute at a tunnel level. However, it is important to note that only one level is taken into account. So, the configuration at the LSP level is considered more specific than the one at the level of the tunnel, and it is used from this point onwards.

Attributes that are not specified within an attribute-set take their values as usual--configuration at the tunnel level, configuration at the global MPLS level, or default values. Here is an example:

attribute-set path-option MYSET
   affinity 0xBEEF mask 0xBEEF

interface tunnel-te 10
   affinity 0xCAFE mask 0xCAFE
   signalled-bandwidth 1000
   path-option 1 dynamic attribute-set name MYSET
   path-option 2 dynamic

In this example, the attribute-set named MYSET is specifying affinity as 0xBEEF. The signalled bandwidth has not been configured in this MYSET . The tunnel 10 , meanwhile, has affinity 0xCAFE configured. LSPs computed from path-option 1 uses the affinity 0xBEEF/0xBEEF, while LSPs computed from path-option 2 uses the affinity 0xCAFE/0xCAFE. All LSPs computed using any of these path-options use signalled-bandwidth as 1000, as this is the only value that is specified only at the tunnel level.

Note	The attributes configured in a path option attribute-set template takes precedence over the same attribute configured under a tunnel. An attribute configured under a tunnel is used only if the equivalent attribute is not specified by the in-use path option attribute-set template.

MPLS traffic engineering allows constraint-based routing (CBR) of IP traffic. One of the constraints satisfied by CBR is the availability of required bandwidth over a selected path. Regular TE tunnel bandwidth is called the global pool . The subpool bandwidth is a portion of the global pool. If it is not in use, the subpool bandwidth is not reserved from the global pool. Therefore, subpool tunnels require a priority higher than that of non-subpool tunnels.

You can configure the signalled-bandwidth path option attribute to use either the global pool (default) or the subpool bandwidth. The signalled-bandwidth value for the path option may be any valid value and the pool does not have to be the same as that which is configured on the tunnel.

Note	When you configure signalled-bandwidth for path options with the signalled-bandwidth bandwidth [sub-pool | global] kbps command, use either all subpool bandwidths or all global-pool bandwidth values.

Reoptimization to a particular path option is not possible if the in-use path option and the new path option do not share the same bandwidth class. The path option switchover operation would fail in such a scenario. Use this command at the EXEC configuration mode to switchover to a newer path option :

mpls traffic-eng switchover tunnel-xx ID path-option index 

Note	Path option switchover between various path options with different bandwidth classes is not allowed.

When path-protection is enabled, a standby LSP is established to protect traffic going over the tunnel. The standby LSP may be established using either the same path option as the primary LSP, or a different one.

The standby LSP is computed to be diverse from the primary LSP, so bandwidth class differences does not matter. This is true in all cases of diversity except node-diversity. With node diversity, it is possible for the standby LSP to share up to two links with the primary LSP, the link exiting the head node, and the link entering the tail node.

If you want to switchover from one path option to another path option and these path options have different classes, the path option switchover is rejected. However, the path option switchover can not be blocked in the path-protection feature. When the standby LSP becomes active using another path option of a different class type, the path option switchover cannot be rejected at the head end. It might get rejected by the downstream node.

Node-diversity is only possible under limited conditions. The conditions that must be met are:

• there is no second path that is both node and link diverse

• the current LSP uses a shared-media link at the head egress or tail ingress

• the shared-media link used by the current LSP permits computation of a node-diverse path

In Cisco IOS XR, reoptimization between different class types would actually be rejected by the next hop. This rejection will occur by an admission failure.

State reporting refers to the PCC sending information to PCEs about the state of LSPs. This is done as state changes occur and is used to keep PCEs informed of changes to the LSP as they occur. State reporting is also used as part of state synchronization and delegation.

Synchronization refers to a procedure that occurs after a PCEP session is established between a PCE and a PCC. The purpose of state synchronization is to download the current LSP database of the PCC to a PCE. This is done through a set of state reports which are marked as synchronizations. This is the first communication to occur after the session is brought up. A full re-send of state reports can also be avoided when the PCE already has an up-to-date version of the LSP database as the version number can be indicated by the PCE during PCEP session establishment.

Delegation is the action by which control of a state is granted to a PCE by the PCC. A PCE to which control was delegated can alter attributes of the LSP. Control is only delegated to one PCE at a time.

• Delegation of control can be revoked from a PCE by the PCC.
• Delegation of control can also be returned to the PCC by the PCE.

State updating refers to the PCE sending information to a PCC to alter the attributes of an LSP. A state update is a message sent by a PCE to a PCC to alter the state of one or more TE tunnels. State updating is allowed only if the PCE has previously been delegated control of the LSP. State updating is also used to return delegated control.

Creation (or instantiation) of an LSP is a procedure by which a PCE instructs a PCC to create an LSP respecting certain attributes. For LSPs created in this manner, the PCE is delegated control automatically. Stateful PCE procedures enable a PCE to instruct a PCC to create a locally sourced tunnel.

Fast repair feature minimizes the tunnel down time by allowing the path computation client (headend) to determine a new optimal path for delegated tunnels that went down, or are under fast reroute (FRR) or soft-preemption. Previously, Path Computation Client (PCC) was not designated to take any action on delegated tunnels. To configure the fast repair feature, use the fast-repair command under PCE stateful client in MPLS-TE configuration.

PCE is still the primary controller, but the time taken to notify the PCE and the wait till the PCE takes an action, amounts to considerable time. This disadvantage is overcome by the fast repair feature.

An optional vendor specific Path Computation Element Protocol (PCEP) extension, cisco-tlv is added in this IOS XR release. The vendor information TLV (Type-Length-Variable) is used to carry vendor specific information that applies to a specific PCEP object by including the TLV in the object.

Vendor specific PCEP extension (cisco-tlv) in not sent in PC report (PCReport), or accepted in PC update (PCUpdate) or PC initiate (PCInitiate) by default, for compatibility reasons. This helps in interoperability with PCE implementation which does not understand or support Cisco specific information.

Vendor specific PCEP extension is optional and can be enabled using the cisco-extension command under PCE stateful client in MPLS-TE configuration.

For detailed steps to enable vendor specific PCEP extension, see Enabling PCEP Cisco Extension.

Automatic bandwidth feature allows a tunnel to automatically and dynamically adjust its reserved bandwidth over time, without network operator intervention. The automatic bandwidth feature support has been extended to delegated tunnels. Previously, tunnels configured with automatic bandwidth were switched to collect-only mode upon delegation.

Affinity is MPLS traffic engineering (TE) tunnel's requirements on the attributes of the links it will cross. The tunnel's affinity bits and affinity mask bits must match the attribute bits of the various links carrying the tunnel.

A new style of affinity reporting support is added in this IOS XR release. Even though TE ignores any affinities from the PCE, the new style affinities in PC update (PCUpdate) or PC initiate (PCInitiate) override the existing tunnel affinities. Previously, only old style affinities (value + mask) were reported. The new affinity mapping has PCEP affinities on the left and IOS XR affinities on the right.

• Lspa.exclude_any = AFFINITY_NEWSTYLE_EXCLUDE OR AFFINITY_NEWSTYLE_EXCLUDE_ALL

• Lspa.include_all = AFFINITY_NEWSTYLE_INCLUDE_STRICT

• Lspa.include_any = AFFINITY_NEWSTYLE_INCLUDE_STRICT OR AFFINITY_NEWSTYLE_INCLUDE

A binding Segment-ID (SID) can be used to enforce traffic engineering (TE) policy using RSVP-TE or SR-TE label switching path (LSP) tunnel. If the topmost label of an incoming packet is the binding SID, the packet is steered to the appropriate LSP tunnel. As such, a SID can be used by an upstream router to steer traffic originating from a downstream router into the appropriate TE path. If an LSP tunnel is PCE controlled, that is, either initiated by PCE or delegated to PCE, or simply reported (without delegation) to a PCE, the router allocates binding label and reports it to the PCE.

Use Case Scenario

A sample use case for binding SID is illustrated in the following diagram.

1. In the MPLS Data Center (DC) network, an SR LSP (without traffic engineering) is established using a prefix SID advertised by BGP.

2. In IP/MPLS WAN, an SR-TE LSP is setup using the PCE. The list of SIDs of the SR-TE LSP is {A, B, C, D}.

3. The gateway node 1 (which is the PCC) allocates a binding SID X and reports it to the PCE.

4. In order for the access node to steer the traffic over the SR-TE LSP, the PCE passes the SID stack {Y, X} where Y is the prefix SID of the gateway node 1 to the access node. In the absence of the binding SID X, the PCE passes the SID stack {Y, A, B, C, D} to the access node.

This example also illustrates the additional benefit of using the binding SID to reduce the number of SIDs imposed on the access nodes with a limited forwarding capacity.

• IGP support is only for IS-IS.

• IS-IS IPv4 and IPv6 must be configured under the same IS-IS instance.

• Unequal load balancing (UELB) does not apply to IPv6 traffic. While it may still be configured and used for IPv4 traffic, IPv6 traffic does not acknowledge the UELB configuration. However, equal loadsharing works for IPv6.

• Policy-based tunnel selection (PBTS) does not apply for IPv6 traffic. While it may still be configured and used for IPv4 traffic, IPv6 traffic does not acknowledge the PBTS configuration.

• MPLS auto tunnels do not support IPv6 autoroute announce and IPv6 forwarding adjacency configurations.

For a given tunnel, you can configure one or more path options - each identified by a unique name. The path option expresses the preference for the path; lower numbers have a higher preference, with 1 having the highest preference. You can also configure the computation method for the path.