RSVP-TE provides support for the communication of exclusion information during Label Switched Path (LSP) setup. A typical LSP diversity use case is for protection, where two LSPs should follow different paths through the network in order to avoid single points of failure, thus greatly improving service availability. This document specifies an approach that can be used for network scenarios where the full path(s) is not necessarily known by use of an abstract identifier for the path. Three types of abstract identifiers are specified: client based, Path Computation Element (PCE) based, and network based. This document specifies two new diversity subobjects for the RSVP eXclude Route Object (XRO) and the Explicit Exclusion Route Subobject (EXRS).¶
For the protection use case, LSPs are typically created at a slow rate and exist for a long time so that it is reasonable to assume that a given (reference) path currently existing (with a well-known identifier) will continue to exist and can be used as a reference when creating the new diverse path. Re-routing of the existing (reference) LSP, before the new path is established, is not considered.¶
This is an Internet Standards Track document.¶
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.¶
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc8390.¶
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
Path diversity for multiple connections is a well-known operational requirement. Diversity constraints ensure that Label Switched Paths (LSPs) can be established without sharing network resources, thus greatly reducing the probability of simultaneous connection failures.¶
The source node can compute diverse paths for LSPs when it has full knowledge of the network topology and is permitted to signal an Explicit Route Object (ERO). However, there are scenarios where different nodes perform path computations, and therefore there is a need for relevant diversity constraints to be signaled to those nodes. These include (but are not limited to):¶
[RFC4874] introduced a means of specifying nodes and resources to be excluded from a route using the eXclude Route Object (XRO) and Explicit Exclusion Route Subobject (EXRS). It facilitates the calculation of diverse paths for LSPs based on known properties of those paths including addresses of links and nodes traversed and Shared Risk Link Groups (SRLGs) of traversed links. Employing these mechanisms requires that the source node that initiates signaling knows the relevant properties of the path(s) from which diversity is desired. However, there are circumstances under which this may not be possible or desirable, including (but not limited to):¶
This is illustrated in Figure 1, where the overlay reference model from [RFC4208] is shown.¶
Overlay Overlay
Network +----------------------------------+ Network
+---------+ | | +---------+
| +----+ | | +-----+ +-----+ +-----+ | | +----+ |
| | | | UNI | | | | | | | | UNI | | | |
| -+ EN1+-+-----+--+ CN1 +----+ CN2 +----+ CN3 +---+-----+-+ EN3+- |
| | | | +--+--+ | | | | | | +---+-| | |
| +----+ | | | +--+--+ +--+--+ +--+--+ | | | +----+ |
+---------+ | | | | | | | +---------+
| | | | | | |
+---------+ | | +--+--+ | +--+--+ | | +---------+
| +----+ | | | | | +-------+ +-----+ | +----+ |
| | +-+--+ | | CN4 +---------------+ CN5 | | | | | |
| -+ EN2+-+-----+--+ | | +---+-----+-+ EN4+- |
| | | | UNI | +-----+ +-----+ | UNI | | | |
| +----+ | | | | +----+ |
+---------+ +----------------------------------+ +---------+
Overlay Core Network Overlay
Network Network
Legend: EN - Edge Node
CN - Core Node
Figure 1 depicts two types of UNI connectivity: single-homed and dual-homed ENs (which also applies to higher-order multihomed connectivity). Single-homed EN devices are connected to a single CN device via a single UNI link. This single UNI link may constitute a single point of failure. UNI connection between EN1 and CN1 is an example of singled-homed UNI connectivity.¶
Such a single point of failure can be avoided when the EN device is connected to two different CN devices, as depicted for EN2 in Figure 1. For the dual-homing case, it is possible to establish two different UNI connections from the same source EN device to the same destination EN device. For example, two connections from EN2 to EN3 may use the two UNI links EN2-CN1 and EN2-CN4. To avoid single points of failure within the provider network, it is necessary to also ensure path (LSP) diversity within the core network.¶
In a network providing a set of UNI interfaces between ENs and CNs such as that shown in Figure 1, the CNs typically perform path computation. Information sharing across the UNI boundary is restricted based on the policy rules imposed by the core network. Typically, the core network topology information as well as LSP path information is not exposed to the ENs. In the network shown in Figure 1, consider a use case where an LSP from EN2 to EN4 needs to be SRLG diverse from an LSP from EN1 to EN3. In this case, EN2 may not know SRLG attributes of the EN1-EN3 LSP and hence cannot construct an XRO to exclude these SRLGs. In this example, EN2 cannot use the procedures described in [RFC4874]. Similarly, an LSP from EN2 to EN3 traversing CN1 needs to be diverse from an LSP from EN2 to EN3 going via CN4. Again, in this case, exclusions based on [RFC4874] cannot be used.¶
This document addresses these diversity requirements by introducing an approach of excluding the path taken by these particular LSP(s). Each reference LSP or route from which diversity is required is identified by an abstract "identifier". The type of identifier to use is highly dependent on the core network operator's networking deployment scenario; it could be client initiated (provided by the EN), provided by a PCE, or allocated by the (core) network. This document defines three different types of identifiers corresponding to these three cases: a client-initiated identifier, a PCE-allocated identifier, and an identifier allocated by the CN ingress node (UNI‑N), i.e., a network-assigned identifier.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here. ¶
The following fields MUST be used to represent the client-initiated identifier: IPv4/IPv6 tunnel sender address, IPv4/IPv6 tunnel endpoint address, Tunnel ID, and Extended Tunnel ID. Based on local policy, the client MAY also include the LSP ID to identify a specific LSP within the tunnel. These fields are defined in Sections 4.6.1.1 and 4.6.2.1 of [RFC3209].¶
The usage of the client-initiated identifier is illustrated by Figure 1. Suppose an LSP from EN2 to EN4 needs to be diverse with respect to an LSP from EN1 to EN3.¶
The LSP identifier of the EN1-EN3 LSP is LSP-IDENTIFIER1, where LSP-IDENTIFIER1 is defined by the tuple ¶
Similarly, the LSP identifier of the EN2-EN4 LSP is LSP-IDENTIFIER2, where LSP-IDENTIFIER2 is defined by the tuple ¶
The EN1-EN3 LSP is signaled with an exclusion requirement from LSP-IDENTIFIER2, and the EN2-EN4 LSP is signaled with an exclusion requirement from LSP-IDENTIFIER1. In order to maintain diversity between these two connections within the core network, the core network SHOULD implement crankback signaling extensions as defined in [RFC4920]. Note that crankback signaling is known to lead to slower setup times and suboptimal paths under some circumstances as described by [RFC4920].¶
In scenarios where a PCE is deployed and used to perform path computation, typically the ingress node of the core network (e.g., node CN1 in Figure 1) could consult a PCE to allocate identifiers, which are used to signal path diversity constraints. In other deployment scenarios, a PCE is deployed at a network node(s) or it is part of a Network Management System (NMS). In all these cases, the PCE is consulted and the Path Key, as defined in [RFC5520], can be used in RSVP signaling as the identifier to ensure diversity.¶
An example of specifying LSP diversity using a Path Key is shown in Figure 2, where a simple network with two domains is shown. It is desired to set up a pair of path-disjoint LSPs from the source in Domain 1 to the destination in Domain 2, but the domains keep strict confidentiality about all path and topology information.¶
The first LSP is signaled by the source with ERO {A, B, loose Dst} and is set up with the path {Src, A, B, U, V, W, Dst}. However, when sending the Record Route Object (RRO) out of Domain 2, node U would normally strip the path and replace it with a loose hop to the destination. With this limited information, the source is unable to include enough detail in the ERO of the second LSP to avoid it taking, for example, the path {Src, C, D, X, V, W, Dst} for path-disjointness.¶
--------------------- -----------------------------
| Domain 1 | | Domain 2 |
| | | |
| --- --- | | --- --- --- |
| | A |--| B |--+--+--| U |--| V |---| W | |
| / --- --- | | --- --- --- \ |
| ---/ | | / / \--- |
| |Src| | | / / |Dst| |
| ---\ | | / / /--- |
| \ --- --- | | --- / --- / --- / |
| | C |--| D |--+--+--| X |---| Y |--| Z | |
| --- --- | | --- --- --- |
| | | |
--------------------- -----------------------------
In order to support LSP diversity, node U consults the PCE and replaces the path segment {U, V, W} in the RRO with a Path Key subobject. The PCE function assigns an "identifier" and puts it into the Path Key field of the Path Key subobject. The PCE ID in the message indicates that this replacement operation was performed by node U.¶
With this additional information, the source node is able to signal the subsequent LSPs with the ERO set to {C, D, exclude Path Key (signaled in the EXRS RSVP subobject), loose Dst}. When the signaling message reaches node X, it can consult the PCE function associated with node U to expand the Path Key in order to calculate a path that is diverse with respect to the first LSP. Alternatively, the source node could use an ERO of {C, D, loose Dst} and include an XRO containing the Path Key.¶
This mechanism can work with all the Path Key resolution mechanisms, as detailed in Section 3.1 of [RFC5553]. A PCE, co-located or not, may be used to resolve the Path Key, but the node (i.e., a Label Switching Router (LSR)) can also use the Path Key information to index a path segment previously supplied to it by the entity that originated the Path Key (for example, the LSR that inserted the Path Key in the RRO or a management system).¶
There are scenarios in which the network provides diversity-related information for a service that allows the client device to include this information in the signaling message. If the Shared Risk Link Group (SRLG) identifier information is both available and shareable (by policy) with the ENs, the procedure defined in [RFC8001] can be used to collect SRLG identifiers associated with an LSP (LSP1). When a second LSP (LSP2) needs to be diverse with respect to LSP1, the EN constructing the RSVP signaling message for setting up LSP2 can insert the SRLG identifiers associated with LSP1 as diversity constraints into the XRO using the procedure described in [RFC4874]. However, if the core network SRLG identifiers are either not available or not shareable with the ENs based on policies enforced by the core network, existing mechanisms cannot be used.¶
In this document, a signaling mechanism is defined where information signaled to the CN via the UNI does not require shared knowledge of core network SRLG information. For this purpose, the concept of a Path Affinity Set (PAS) is defined for abstracting SRLG information. The motive behind the introduction of the PAS is to minimize the exchange of diversity information between the core network (CNs) and the client devices (ENs). The PAS contains an abstract SRLG identifier associated with a given path rather than a detailed SRLG list. The PAS is a single identifier that can be used to request diversity and associate diversity. The means by which the processing node determines the path corresponding to the PAS is beyond the scope of this document.¶
A CN on the core network boundary interprets the specific PAS identifier (e.g., "123") as meaning to exclude the core network SRLG information (or equivalent) that has been allocated by LSPs associated with this PAS identifier value. For example, if a path exists for the LSP with the PAS identifier "123", the CN would use local knowledge of the core network SRLGs associated with the LSPs tagged with PAS attribute "123" and use those SRLGs as constraints for path computation. If a PAS identifier is used as an exclusion identifier in the connection request, the CN (UNI-N) in the core network is assumed to be able to determine the existing core network SRLG information and calculate a path that meets the determined diversity constraints.¶
When a CN satisfies a connection setup for an SRLG-diverse signaled path, the CN may optionally record the core network SRLG information for that connection in terms of CN-based parameters and associate that with the EN addresses in the Path message. Specifically, for Layer 1 Virtual Private Networks (L1VPNs), Port Information Tables (PITs) [RFC5251] can be leveraged to translate between client (EN) addresses and core network addresses.¶
The means to distribute the PAS information within the core network is beyond the scope of this document. For example, the PAS and the associated SRLG information can be distributed within the core network by an Interior Gateway Protocol (IGP) or by other means such as configuration. Regardless of means used to distribute the PAS information, the information is kept inside the core network and is not shared with the overlay network (see Figure 1).¶
This section describes the signaling extensions required to address the aforementioned requirements and use cases.¶
New Diversity XRO subobjects are defined below for the IPv4 and IPv6 address families. Most of the fields in the IPv4 and IPv6 Diversity XRO subobjects are common and are described following the definition of the two subobjects.¶
The IPv4 Diversity XRO subobject is defined as follows:¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|L| XRO Type | Length |DI Type|A-Flags|E-Flags| Resvd |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv4 Diversity Identifier Source Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Diversity Identifier Value |
// ... //
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
Similarly, the IPv6 Diversity XRO subobject is defined as follows:¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|L| XRO Type | Length |DI Type|A-Flags|E-Flags| Resvd |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Diversity Identifier Source Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Diversity Identifier Source Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Diversity Identifier Source Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Diversity Identifier Source Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Diversity Identifier Value |
// ... //
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
The L flag is used in the same way as for the XRO subobjects defined in [RFC4874], that is: ¶
0 indicates that the diversity constraints MUST be satisfied, and ¶
1 indicates that the diversity constraints SHOULD be satisfied. ¶
The value is set to 38 for the IPv4 Diversity XRO subobject. The value is set to 39 for the IPv6 Diversity XRO subobject. ¶
Per [RFC4874], the Length contains the total length of the IPv4/IPv6 subobject in bytes, including the XRO Type and Length fields. The Length is variable, depending on the Diversity Identifier Value. ¶
Diversity Identifier Type (DI Type) indicates the way the reference LSP(s) or route(s) with which diversity is required is identified in the IPv4/IPv6 Diversity subobjects. The following three DI Type values are defined in this document: ¶
DI Type value Definition
------------- --------------------------------
1 Client-Initiated Identifier
2 PCE-Allocated Identifier
3 Network-Assigned Identifier
¶
The Attribute Flags (A-Flags) are used to communicate desirable attributes of the LSP being signaled in the IPv4/IPv6 Diversity subobjects. Each flag acts independently. Any combination of flags is permitted. ¶
Indicates that the exclusion does not apply to the destination node of the LSP being signaled. ¶
Indicates that the exclusion does not apply to the node(s) performing ERO expansion for the LSP being signaled. An ingress UNI-N node is an example of such a node. ¶
Indicates that the penultimate node of the LSP being signaled MAY be shared with the excluded path even when this violates the exclusion flags. This flag is useful, for example, when an EN is not dual homed (like EN4 in Figure 1, where all LSPs have to go through CN5). ¶
The "Penultimate node exception" flag is typically set when the destination node is single homed (e.g., EN1 or EN4 in Figure 2). In such a case, LSP diversity can only be accomplished inside the core network up to the egress node and the penultimate hop must be the same for the LSPs. ¶
This flag is used to indicate tunnel-level exclusion. Specifically, this flag is used to indicate that if the diversity identifier contains an LSP ID field, then the LSP ID is to be ignored, and the exclusion applies to any LSP matching the rest of the diversity identifier.¶
The Exclusion Flags are used to communicate the desired type(s) of exclusion requested in the IPv4/IPv6 Diversity subobjects. The following flags are defined. Any combination of these flags is permitted. Please note that the exclusion specified by these flags may be modified by the value of the A-Flags. For example, the node exclusion flag is ignored for the penultimate node if the "Penultimate node exception" flag of the A-Flags is set. ¶
Indicates that the path of the LSP being signaled is requested to be SRLG disjoint with respect to the excluded path specified by the IPv4/IPv6 Diversity XRO subobject.¶
Indicates that the path of the LSP being signaled is requested to be "node diverse" from the excluded path specified by the IPv4/IPv6 Diversity XRO subobject.¶
Indicates that the path of the LSP being signaled is requested to be "link diverse" from the path specified by the IPv4/IPv6 Diversity XRO subobject.¶
This flag is reserved. It MUST be set to zero on transmission and MUST be ignored on receipt for both IPv4/IPv6 Diversity XRO subobjects.¶
This field is reserved. It MUST be set to zero on transmission and MUST be ignored on receipt for both IPv4/IPv6 Diversity XRO subobjects. ¶
This field MUST be set to the IPv4/IPv6 address of the node that assigns the diversity identifier. Depending on the Diversity Identifier Type, the diversity identifier source may be a client node, PCE entity, or network node. Specifically: ¶
Encoding for this field depends on the Diversity Identifier Type, as defined in the following. ¶
When the Diversity Identifier Type is set to "Client-Initiated Identifier" in the IPv4 Diversity XRO subobject, the Diversity Identifier Value MUST be encoded as follows: ¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv4 Tunnel Endpoint Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Must Be Zero | Tunnel ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Extended Tunnel ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Must Be Zero | LSP ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Tunnel Endpoint Address |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Tunnel Endpoint Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Tunnel Endpoint Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| IPv6 Tunnel Endpoint Address (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Must Be Zero | Tunnel ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Extended Tunnel ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Extended Tunnel ID (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Extended Tunnel ID (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Extended Tunnel ID (cont.) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Must Be Zero | LSP ID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Must Be Zero | Path Key |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Path Affinity Set (PAS) Identifier |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
[RFC4874] defines the EXRS ERO subobject. An EXRS is used to identify abstract nodes or resources that must not or should not be used on the path between two inclusive abstract nodes or resources in the explicit route. An EXRS contains one or more subobjects of its own, called EXRS subobjects [RFC4874].¶
An EXRS MAY include a Diversity subobject as specified in this document. The same type values 38 and 39 MUST be used.¶
The procedure defined in [RFC4874] for processing the XRO and EXRS is not changed by this document. The processing rules for the Diversity XRO and EXRS subobjects are similar unless the differences are explicitly described. Similarly, IPv4 and IPv6 Diversity XRO subobjects and IPv4 and IPv6 Diversity EXRS subobjects follow the same processing rules.¶
If the processing node cannot recognize the Diversity XRO/EXRS subobject, the node is expected to follow the procedure defined in [RFC4874].¶
An XRO/EXRS object MAY contain multiple Diversity subobjects of the same DI Type. For example, in order to exclude multiple Path Keys, a node MAY include multiple Diversity XRO subobjects, each with a different Path Key. Similarly, in order to exclude the routes taken by multiple LSPs, a node MAY include multiple Diversity XRO/EXRS subobjects, each with a different LSP identifier. Likewise, to exclude multiple PAS identifiers, a node MAY include multiple Diversity XRO/EXRS subobjects, each with a different PAS identifier. However, all Diversity subobjects in an XRO/EXRS MUST contain the same Diversity Identifier Type. If a Path message contains an XRO/EXRS with multiple Diversity subobjects of different DI Types, the processing node MUST return a PathErr with the error code "Routing Problem" (24) and error sub-code "XRO/EXRS Too Complex" (68/69).¶
If the processing node recognizes the Diversity XRO/EXRS subobject but does not support the DI Type, it MUST return a PathErr with the error code "Routing Problem" (24) and error sub-code "Unsupported Diversity Identifier Type" (36).¶
In the case of DI Type "Client-Initiated Identifier", all nodes along the path SHOULD process the diversity information signaled in the XRO/EXRS Diversity subobjects to verify that the signaled diversity constraint is satisfied. If a diversity violation is detected, crankback signaling MAY be initiated.¶
In the case of DI Type "PCE-Allocated Identifier" and "Network-Assigned Identifier", the nodes in the domain that perform path computation SHOULD process the diversity information signaled in the XRO/EXRS Diversity subobjects as follows. In the PCE case, the ingress node of a domain sends a path computation request for a path from ingress node to egress node, including diversity constraints to a PCE. Or, in the PAS case, the ingress node is capable of calculating the path for the new LSP from ingress node to the egress node, taking the diversity constraints into account. The calculated path is then carried in the Explicit Route Object (ERO). Hence, the transit nodes in a domain and the domain egress node SHOULD NOT process the signaled diversity information unless path computation is performed.¶
While processing the EXRS object, if a loose hop expansion results in the creation of another loose hop in the outgoing ERO, the processing node MAY include the EXRS in the newly created loose hop for further processing by downstream nodes.¶
The A-Flags affect the processing of the Diversity XRO/EXRS subobject as follows: ¶
When the "Penultimate node exception" flag is set in the XRO subobject, the exclusion MUST be ignored for the penultimate node on the path of the LSP being established. ¶
The penultimate node exception for the EXRS subobject applies to the node before the explicit node identified by the ERO subobject that identifies the next abstract node. When the "Penultimate node exception" flag is set in the EXRS subobject, the exclusion MUST be ignored for said node (i.e., the node before the next abstract node). ¶
If the L-flag of the Diversity XRO subobject or Diversity EXRS subobject is not set, the processing node proceeds as follows.¶
If the L-flag of the Diversity XRO subobject or Diversity EXRS subobject is set, the processing node proceeds as follows:¶
If, subsequent to the initial signaling of a diverse LSP, an excluded path referenced in the XRO subobject becomes known to the processing node or a change in the excluded path becomes known to the processing node, the processing node MUST re-evaluate the exclusion and diversity constraints requested by the diverse LSP to determine whether they are still satisfied. ¶
This document does not introduce any additional security issues in addition to those identified in [RFC5920], [RFC2205], [RFC3209], [RFC3473], [RFC2747], [RFC4874], [RFC5520], and [RFC5553].¶
The diversity mechanisms defined in this document rely on the new diversity subobject that is carried in the XRO or EXRS, respectively. In Section 7 of [RFC4874], it is noted that some administrative boundaries may remove the XRO due to security concerns on explicit route information exchange. However, when the diversity subobjects specified in this document are used, removing at the administrative boundary an XRO containing these diversity subobjects would result in the request for diversity being dropped at the boundary, and path computation would be unlikely to produce the requested diverse path. As such, diversity subobjects MUST be retained in an XRO crossing an administrative boundary, even if other subobjects are removed. This retention would be based on operator policy. The use of diversity subobjects is based on mutual agreement. This avoids the need to share the identity of network resources when supporting diversity.¶
IANA has assigned new values defined in this document and summarized in this section.¶
In the IANA registry for RSVP parameters, under "Class Names, Class Numbers, and Class Types", this document defines two new subobjects for the EXCLUDE_ROUTE object [RFC4874], C-Type 1 (see "Class Types or C-Types - 232 EXCLUDE_ROUTE" on <https://www.iana.org/assignments/rsvp-parameters>).¶
| Description | Value |
|---|---|
| IPv4 Diversity | 38 |
| IPv6 Diversity | 39 |
The Diversity XRO subobjects are also defined as new EXRS subobjects (see "Class Types or C-Types - 20 EXPLICIT_ROUTE" on <https://www.iana.org/assignments/rsvp-parameters>). The same numeric values have been assigned:¶
| Description | Value |
|---|---|
| IPv4 Diversity | 38 |
| IPv6 Diversity | 39 |
In the IANA registry for RSVP parameters, under "Error Codes and Globally Defined Error Value Sub-Codes", for Error Code "Routing Problem" (24) (see [RFC3209]), the following sub-codes are defined (see "Sub-Codes - 24 Routing Problem" on <https://www.iana.org/assignments/rsvp-parameters>).¶
| Value | Description | Reference |
|---|---|---|
| 36 | Unsupported Diversity Identifier Type | RFC 8390 |
For Error Code "Notify Error" (25) (see [RFC3209]), the following sub-codes are defined (see "Sub-Codes - 25 Notify Error" on <https://www.iana.org/assignments/rsvp-parameters>).¶
| Value | Description | Reference |
|---|---|---|
| 14 | Route of XRO LSP identifier unknown | RFC 8390 |
| 15 | Failed to satisfy Exclude Route | RFC 8390 |
| 16 | Compliant path exists | RFC 8390 |
The authors would like to thank Xihua Fu for his contributions. The authors also would like to thank Luyuan Fang and Walid Wakim for their review and comments.¶
Igor Bryskin Huawei Technologies Email: Igor.Bryskin@huawei.com Daniele Ceccarelli Ericsson Email: Daniele.Ceccarelli@ericsson.com Dhruv Dhody Huawei Technologies Email: dhruv.ietf@gmail.com Don Fedyk Hewlett-Packard Enterprise Email: don.fedyk@hpe.com Clarence Filsfils Cisco Systems, Inc. Email: cfilsfil@cisco.com Gabriele Maria Galimberti Cisco Systems Email: ggalimbe@cisco.com Ori Gerstel SDN Solutions Ltd. Email: origerstel@gmail.com Oscar Gonzalez de Dios Telefonica I+D Email: ogondio@tid.es Matt Hartley Cisco Systems Email: mhartley@cisco.com Kenji Kumaki KDDI Corporation Email: ke-kumaki@kddi.com Ruediger Kunze Deutsche Telekom AG Email: Ruediger.Kunze@telekom.de Lieven Levrau Nokia Email: Lieven.Levrau@nokia.com Cyril Margaria Email: cyril.margaria@gmail.com Julien Meuric France Telecom Orange Email: julien.meuric@orange.com Yuji Tochio Fujitsu Email: tochio@jp.fujitsu.com Xian Zhang Huawei Technologies Email: zhang.xian@huawei.com¶