]> Huawei

italo.busi@huawei.com

Alef Edge

xufeng.liu.ietf@gmail.com

Individual

i_bryskin@yahoo.com

Cisco Systems Inc

tsaad.net@gmail.com

Telefonica

oscar.gonzalezdedios@telefonica.com

TEAS Working Group This document describes how profiles of the Topology YANG data model, defined in RFC8795, can be used to address applications in Traffic Engineering aware (TE-aware) deployments, irrespective of whether they are TE-centric or not.

Introduction Many network scenarios are being discussed in various IETF Working Groups (WGs) that are not classified as "Traffic Engineering" use cases but can be addressed by a profile (sub-set) of the Topology YANG data model, defined in . Traffic Engineering (TE) is defined in as aspects of Internet network engineering that deal with the issues of performance evaluation and performance optimization of operational IP networks. TE encompasses the application of technology and scientific principles to the measurement, characterization, modeling, and control of Internet traffic. The Topology YANG data model, defined in , augments the Network Topology YANG data model, defined in , with generic and technology-agnostic features that are not only applicable to TE-centric deployments, but also applicable to non-TE-centric yet TE-aware deployments. A TE-aware deployment is one where the topology carries information that can be used to influence how traffic can be engineered within the network. In some scenarios, this information can be leveraged even in use cases where traffic doesn't need to be engineered. Examples of generic TE-aware features that can apply to both TE-centric and non-TE-centric use-cases are: inter-domain link discovery (plug-id), geo-localization, multi-layer topology representation, node-specific switching limitation, link reliability, and topology abstraction. It is also worth noting that also the boundary between the TE-specific model constructs and the core network topology model constructs is also blurred since new applications may need to leverage on constructs which was originally defined to support TE-centric scenarios but that are also needed to support these new applications. An example of a concept that originated from TE-centric scenarios but can be considered a core network topology model construct is the SRLG. New applications such as what-if analysis need to be aware of the SRLG information also for non-TE-centric scenarios to provide reliable results. It is also worth noting that the Topology YANG data model, defined in , is quite an extensive and comprehensive model in which most features are optional. Therefore, even though the full model appears to be complex, at the first glance, a profile (sub-set) of the model can be used to address specific scenarios irrespective of whether they are TE-centric or not. The implementation of profiles can simplify and expedite adoption of the Topology YANG data model, defined , and allow for its reuse even for non-TE-centric use-cases. The key question is whether all or some of the attributes defined in the Topology YANG data model, defined in , are needed to address a given network scenario. provides examples where profiles of the Topology YANG data model, defined in , can be used to address some generic use cases applicable to both TE-centric and non-TE-centric deployments. Understanding that these profiles are generic would be more straightforward if the profiled YANG data nodes where defined under a container with a different name than 'te' or directly under the parent YANG data node. However, the 'te' container in the context of , should be understood as the container of YANG data nodes providing TE-aware topology information.

Examples of generic profiles

UNI Topology Discovery The following profile of the Topology YANG data model, defined in , can be used to support the UNI Topology Discovery, or in general, inter-domain link discovery:

It is also worth noting that the UNI links can also be TE (e.g. an OTN UNI) or non-TE (e.g., an Ethernet UNI) as well as multi-function UNI links, supporting both TE and non-TE technologies, such as the UNI links, described in , which can be configured either OTN UNI or Ethernet UNI or SDH UNI. The UNI Topology profiled YANG data model shown in can also be used with technology-specific UNI augmentations, as described in . Technology-specific augmentations can for example describe the capability of the TP to be configured as a UNI for the types of services supported by the UNI (e.g., L2VPN/L3VPN). For example, in , the eth-svc container is defined to represent the capabilities of the Termination Point (TP) to be configured as an Ethernet UNI, together with the Ethernet classification and VLAN operations supported by that TP. The provides another example, where: the client-svc container is defined to represent the capabilities of the TP to be configured as an transparent client UNI (e.g., STM-N, Fiber Channel or transparent Ethernet); the OTN technology-specific Link Termination Point (LTP) augmentations are defined to represent the capabilities of the TP to be configured as an OTN UNI, together with the information about OTN label and bandwidth availability at the OTN UNI. Te UNI Topology profiled YANG data model shown in does not require the network to be a TE network and, therefore, it could be used as a core network topology model to discover UNIs (or in general any external link) for TE and non-TE networks as well as multi-layer networks encompassing both TE and non-TE layers. The advantages of using the UNI Topology profiled YANG data model shown in as a core network topology model is to have a common solutions for: discovering UNIs as well as inter-domain NNI links, which is applicable to any technology (TE or non TE) used at the UNI or within the network; modelling non TE UNIs such as Ethernet, and TE UNIs such as OTN, as well as UNIs which can configured as TE or non-TE (e.g., being configured as either Ethernet or OTN UNI).

Administrative and Operational status management The following profile of the Topology YANG data model, defined in , can be used to manage the administrative and operational for nodes, termination points and links as well as to associate some administrative names to network topologies, nodes, termination points and links:

Overlay and Underlay Topologies The following profile of the Topology YANG data model, defined in , can be used to manage the overlay/underlay relationships for nodes and links, as described in section 5.8 of :

/nw:networks/network/network-id augment /nw:networks/nw:network/nt:link: +--rw te! +--rw te-link-attributes +--rw underlay {te-topology-hierarchy}? +--rw enabled? boolean +--rw primary-path +--rw network-ref? | -> /nw:networks/network/network-id +--rw path-element* [path-element-id] +--rw path-element-id uint32 +--rw (type)? +--:(numbered-link-hop) | +--rw numbered-link-hop | +--rw link-tp-id te-tp-id | +--rw hop-type? te-hop-type | +--rw direction? te-link-direction +--:(unnumbered-link-hop) +--rw unnumbered-link-hop +--rw link-tp-id te-tp-id +--rw node-id te-node-id +--rw hop-type? te-hop-type +--rw direction? te-link-direction ]]>

The advantages of using the underlay/overlay network profiled YANG data model shown in as a core network topology model is to have a common solutions for navigating between overlay/underlay network topologies where: both the underlay and overlay network topologies are TE networks; both the underlay and overlay network topologies are non-TE networks; the underlay and overlay network topologies are TE and non-TE networks; the underlay or the overlay network topology is a multi-layer network encompassing both TE and non-TE layers.

Supporting relationships in RFC8345 defines the modeling constructs for supporting relations, including supporting network (i.e. topology), supporting node, supporting link, and supporting termination point. These relation constructs facilitate network mappings and transformations. One use case is to map a customized topology to a native topology. The customized topology uses different name spaces from the native topology when naming nodes, links, or termination points. There is a supporting relationship between a modeling construct in the customized topography to its counterpart in the native topology. In such a relationship, the modeling constructs at both ends represent the same type of network objects, which can be network (i.e. topology), node, link, or termination point. defines the modeling constructs for network overlay and underlay relations. When the network overlay technology is used, some network objects (nodes or links) in the overlay network are built on top of network objects in the underlay network. As a result, the overlay-underlay relationship is created between network objects in the overlay networks and other network objects in the underlay network. Between the network object pairs in the overlay-underlay relationship, the types of the network objects are usually not the same. The network object can be a node in the overlay network, while the related underlay network object is a topology in the underlay network. A link in the overlay network can be related to a path that consists of a sequence of nodes and links in the underlay network.

Nodes with switching limitations It is worth noting that a node, as defined in , does not provide any information about the possible connectivity between its TPs. A node can have some switching limitations where connectivity is not possible between all its TP pairs, for example when: the node represents a physical device with switching limitations; the node represents an abstraction of a network topology. The following profile of the Topology YANG data model, defined in , can be used for the management of nodes with switching limitations by defining the node connectivity matrix to represent feasible connections between termination points across the nodes:

Multipoint links According to , multipoint links can be "represented through pseudonodes (similar to IS-IS, for example)". Each access point can have different directionality with respect to the multipoint link, as shown in : - an access point of a multipoint link can be able to both transmit and receive traffic: this access point can be modelled as a TP (e.g., TP A in ) terminating two links, one incoming link (e.g., Link 1 in ) and one outgoing link (e.g., Link 2 in ); - an access point of a multipoint link can be able only to receive traffic: this access point can be modelled as a TP (e.g., TP B in ) with only one incoming link (e.g., Link 3 in ); - an access point of a multipoint link can be able only to transmit traffic: this access point can be modelled as a TP (e.g., TP C in ) with only one outgoing link (e.g., Link 4 in ).

/ \| |/ \ | A | Psedonode | C |-----> <-----\ /| |\ / +-+ | | +-+ Link 4 Link 2 | | | | | | \ / \ / +-------------------+ ]]>

The switching limitations of the pseudonode, as defined in , provides sufficient information to identify the type of multipoint link: - in case of multipoint links, the connectivity matrix of the pseudnode, reports that connectivity is enabled by default between all the TPs of the node; - in case of point-to-multipoint links, the connectivity matrix of the pseudnode, reports that connectivity is possible only between the root TP and the leaf TPs >> - if the point-to-multipoint link is bidirectional, the connectivity matrix of the pseudonodes reports that connectivity is possible from the root TP to the leaf TPs as well as from the leaf TPs to the root TP; >> - the connectivity matrix of the psuedonode can also describe point-to-multipoint links with more than one root (also known as rooted-multipoint links), indicating also whether connectivity between root TPs is allowed or not; - in case of hybrid multipoint links, as defined in , the connectivity matrix of the pseunode reports the list of TP pairs for which connectivity is allowed or not allowed. It is worth noting that the directionality of the access point of a multipoint link overrides the switching limitations of the pseudonode. For example, even if the connectivity matrix of the psuedonode in indicates that connectivity is possible between TP A and TP B, the traffic entering the pseudonode from TP A cannot be transmitted by TP B since there is no outgoing link from TP B. Therefore, the connectivity matrix of a pseudonode modelling a point-to-multipoint unidirectional link, does not need to report that connectivity is only possible from the root TP to the leaf TPs but it can report that connectivity is possible by default between all the TPs of the node. The pseudonode represents a point-to-multipoint unidirectional link, as indicated by a single root TP that can only receive traffic and one or more leaf TPs that can only transmit traffic.

\ /| |\ / +-+ | | +-+ | | | | | | \ / \ / +-------------------+ ]]>

For example, shows an example of a pseudonode representing an unidirectional point-to-multipoint link where the TP A is the root TP and TPs B and C are the two leaf TPs.

Technology-specific augmentations There are two main options to define technology-specific Topology Models which can use the attributes defined in the Topology YANG data model, defined in . Both options are applicable to any possible profile of the TE Topology Model, such as those defined in . The first option is to define a technology-specific TE Topology Model which augments the TE Topology Model, as shown in :

This approach is more suitable for cases when the technology-specific TE topology model provides augmentations to the TE Topology constructs, such as bandwidth information (e.g., link bandwidth), tunnel termination points (TTPs) or connectivity matrices. It also allows providing augmentations to the Network Topology constructs, such as nodes, links, and termination points (TPs). This is the approach currently used in and . It is worth noting that a profile of the technology-specific TE Topology model not using any TE topology attribute or constructs can be used to address any use case that do not require these attributes. In this case, only the 'te-topology' presence container of the TE Topology Model needs to be implemented because it is the parent container for the 'network-type' representing the technology-specific topology model. Other data nodes defined in the TE Topology Model do not need to be implemented by this profile. The second option is to define a technology-specific Network Topology Model which augments the Network Topology Model and to rely on the multiple inheritance capability, which is implicit in the network- types definition of , to allow using also the generic attributes defined in the TE Topology model:

This approach is more suitable in cases where the technology-specific Network Topology Model provides augmentation only to the constructs defined in the Network Topology Model, such as nodes, links, and termination points (TPs). Therefore, with this approach, only the generic attributes defined in the TE Topology Model could be used. It is also worth noting that in this case, technology-specific augmentations for the bandwidth information could not be defined. In principle, it would be also possible to define both a technology specific TE Topology Model which augments the TE Topology Model, and a technology-specific Network Topology Model which augments the Network Topology Model and to rely on the multiple inheritance capability, as shown in :

This option does not provide any technical advantage with respect to the first option, shown in , but could be useful to add augmentations to the TE Topology constructs and to re-use an already existing technology-specific Network Topology Model. It is worth noting that the technology-specific TE Topology model can reference constructs defined by the technology-specific Network Topology model but it could not augment constructs defined by the technology-specific Network Topology model.

Multi-inheritance As described in section 4.1 of , the network types should be defined using presence containers to allow the representation of network subtypes. The hierarchy of network subtypes can be single hierarchy, as shown in . In this case, each presence container contains at most one child presence container, as shows in the JSON code below:

The hierarchy of network subtypes can also be multi-hierarchy, as shown in and . In this case, one presence container can contain more than one child presence containers, as show in the JSON codes below:

Other examples of multi-hierarchy topologies are described in .

Example (Link augmentation) This section provides an example on how technology-specific attributes can be added to the Link construct:

../../../nw:node/node-id | +--rw source-tp? leafref +--rw destination | +--rw dest-node? -> ../../../nw:node/node-id | +--rw dest-tp? leafref +--rw supporting-link* [network-ref link-ref] | +--rw network-ref | | -> ../../../nw:supporting-network/network-ref | +--rw link-ref leafref +--rw example-link-attributes | <...> +--rw te! +--rw te-link-attributes +--rw name? string +--rw example-te-link-attributes | <...> +--rw max-link-bandwidth +--rw te-bandwidth +--rw (technology)? +--:(generic) | +--rw generic? te-bandwidth +--:(example) +--rw example? example-bandwidth ]]>

The technology-specific attributes within the example-link-attributes container can be defined either in the technology-specific TE Topology Model (Option 1) or in the technology-specific Network Topology Model (Option 2 or Option 3). These attributes can only be non-TE and do not require the implementation of the te container. The technology-specific attributes within the example-te-link-attributes container as well as the example max-link-bandwidth can only be defined in the technology-specific TE Topology Model (Option 1 or Option 3). These attributes can be TE or non-TE and require the implementation of the te container.

Open Issues

Implemented profiles When a server implements a profile of the TE topology model, there is no standardized mechanism for the server to report to the client the subset of the model being implemented. This might not be an issue in case the TE topology profile is read by the the client because the server reports in the operational datastore only the leaves which have been implemented, as described in section 5.3 of . More investigation is instead required in case the TE topology profile is configured by the client, to avoid that the client tries to write an attribute not used in the TE Topology profile implemented by the server. It is also worth noting that the supported profile may also depend on other attributes (for example the network type), so the YANG deviation mechanism is not applicable to this scenario. Note: that this issue is also tracked in github as issue #161.

Security Considerations This document provides only information about how the Topology YANG data model, defined in , can be profiled to address some scenarios which are not considered as TE. As such, this document does not introduce any additional security considerations besides those already defined in .

IANA Considerations This document requires no IANA actions.

Acknowledgments The authors would like to thank Vishnu Pavan Beeram, Daniele Ceccarelli, Jonas Ahlberg and Scott Mansfield for providing useful suggestions for this draft. This document was prepared using kramdown. Initial versions of this document were prepared using 2-Word-v2.0.template.dot.

This document defines a YANG data model for representing, retrieving, and manipulating Traffic Engineering (TE) Topologies. The model serves as a base model that other technology-specific TE topology models can augment. This document defines an abstract (generic, or base) YANG data model for network/service topologies and inventories. The data model serves as a base model that is augmented with technology-specific details in other, more specific topology and inventory data models. Datastores are a fundamental concept binding the data models written in the YANG data modeling language to network management protocols such as the Network Configuration Protocol (NETCONF) and RESTCONF. This document defines an architectural framework for datastores based on the experience gained with the initial simpler model, addressing requirements that were not well supported in the initial model. This document updates RFC 7950. Old Dog Consulting This document describes the principles of traffic engineering (TE) in the Internet. The document is intended to promote better understanding of the issues surrounding traffic engineering in IP networks and the networks that support IP networking, and to provide a common basis for the development of traffic engineering capabilities for the Internet. The principles, architectures, and methodologies for performance evaluation and performance optimization of operational networks are also discussed. This work was first published as RFC 3272 in May 2002. This document obsoletes RFC 3272 by making a complete update to bring the text in line with best current practices for Internet traffic engineering and to include references to the latest relevant work in the IETF. Huawei Old Dog Consulting Huawei CAICT This document provides an analysis of the applicability of the YANG models defined by the IETF (in particular in the Traffic Engineering Architecture and Signaling (TEAS) and Common Control and Measurement Plane (CCAMP) working groups) to support ODU transit services, transparent client services, and Ethernet Private Line/Ethernet Virtual Private Line (EPL/EVPL) services over Optical Transport Network (OTN) in single and multi-domain network scenarios. This document also describes how existing YANG models can be used through several worked examples and JSON fragments. Huawei Technologies Huawei Technologies Futurewei Huawei Technologies CAICT China Mobile Alef Edge This document describes a YANG data model for Ethernet networks when used either as a client-layer network of an underlay transport network (e.g., an Optical Transport Network (OTN)) or as a transport network itself. Huawei Technologies Huawei Technologies Alef Edge Nokia Telefonica This document defines a YANG data model for representing, retrieving, and manipulating Optical Transport Network (OTN) topologies. It is independent of control plane protocols and captures topological and resource-related information pertaining to OTN. Huawei Everything OPS Telefonica Swisscom This document defines the concept of Service & Infrastructure Maps (SIMAP) and identifies a set of SIMAP requirements and use cases. The SIMAP was previously known as Digital Map. The document intends to be used as a reference for the assessment of the various topology modules to meet SIMAP requirements. Alef Edge Individual Juniper Networks Juniper Networks Ciena Cisco This document defines a YANG data model for Segment Routing (SR) topology and Segment Routing (SR) traffic engineering (TE) topology, using MPLS data plane.

Contributors Futurewei Inc.

aihuaguo.ietf@gmail.com

Huawei

zhenghaomian@huawei.com

Nokia

sergio.belotti@nokia.com