YANG models for Virtual Network (VN)/TE Performance Monitoring Telemetry and Scaling Intent Autonomics

Introduction The YANG model in is used to operate customer-driven Virtual Networks (VNs) during the computation of VN, its instantiation, and its life-cycle service management and operations. YANG model in is used to operate TE-tunnels during the tunnel instantiation, and its life-cycle management and operations. The models presented in this draft allow the applications hosted by the customers to subscribe to and monitor their key performance data of their interest on the level of VN or TE-tunnel . The key characteristic of the models presented in this document is a top-down programmability that allows the applications hosted by the customers to subscribe to and monitor key performance data of their interest and autonomic scaling intent mechanism on the level of VN as well as TE-tunnel. According to the classification of , the YANG data models presented in this document can be classified as customer service models. These can be mapped to the CMI (Customer Network Controller (CNC)- Multi-Domain Service Coordinator (MSDC) interface) of ACTN . describes key network performance data to be considered for end-to-end path computation in TE networks. The services provided can be optimized to meet the requirements (such as traffic patterns, quality, and reliability) of the applications hosted by the customers. This document provides YANG data models with performance monitoring parameters that can be subscribed to for monitoring and telemetry for any VN/TE-Tunnel via mechanism specified in and . It also provides an ability to program their customized automatic scaling in/out intent as well. These models can be utilized by a client network controller to initiate the capabilities via a NETCONF or a RESTCONF interface. The term 'Performance monitoring' in this document refers to subscription and publication of streaming telemetry data. Subscription is initiated by the client (e.g., CNC) while publication is provided by the network (e.g., MDSC/Provisioning Network Controller (PNC)) based on the client's subscription. As per , this would be classified as a passive method. Note that the actual measurements might be done via any technique though. As the scope of performance monitoring in this document is augment the performance monitoring parameters (telemetry data) on the level of a client's VN or TE-tunnel, the entity interfacing to the client (e.g., MDSC) has to provide VN or TE-tunnel level information. This requires the controller to have the capability to derive VN or TE-tunnel level performance data based on lower-level data collected via PM counters in the Network Elements (NE). How the controller entity derives such customized level data (i.e., VN or TE-tunnel level) is out of the scope of this document. The data model includes configuration and state data according to the Network Management Datastore Architecture (NMDA) .

Terminology Refer to , , and for the key terms used in this document. Scaling: This refers to the network's ability to re-shape its own resources. "Scale out" refers to improve network performance by increasing the allocated resources, while "scale in" refers to decreasing the allocated resources, typically because the existing resources are unnecessary. Scaling Intent: Scaling intent is used to declare scaling conditions. Specifically, scaling intent refers to how the client programs or configures conditions that will be applied to their key performance data to trigger either scaling out or scaling in. Various conditions can be set for scaling intent on either VN or TE-tunnel level. Network Autonomics: This refers to the network automation capability that allows a client to initiate scaling intent mechanisms and provides the client with the status of the adjusted network resources based on the client's scaling intent in an automated fashion.

Tree Diagram A simplified graphical representation of the data model is used in and of this document. The meaning of the symbols in these diagrams is defined in .

Prefixes in Data Node Names In this document, names of data nodes and other data model objects are prefixed using the standard prefix associated with the corresponding YANG imported modules, as shown in Table 1. Prefixes and corresponding YANG modules

Prefix	YANG module	Reference
te	ietf-te
te-types	ietf-te-types
rt-types	ietf-routing-types
te-tel	ietf-te-telemetry	[RFCXXXX]
vn	ietf-vn
vn-tel	ietf-vn-telemetry	[RFCXXXX]

Note: The RFC Editor is requested to replace XXXX with the number assigned to the RFC once this draft becomes an RFC, and to remove this note. Further, the following additional documents are referenced in the model defined in this document -

- OSPF Traffic Engineering (TE) Metric Extensions.
- IS-IS Traffic Engineering (TE) Metric Extensions.
- Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions.

Use-Cases There is a need for real-time (or semi-real-time) traffic monitoring of the network to optimize the network and the traffic distribution. shows the an example of a high-level work-flow for dynamic service control based on traffic monitoring that could use the mechanism described in this document.

Work flow for dynamic service control based on traffic monitoring Some of the key points are as follows:

Network traffic monitoring is important to facilitate automatic discovery of the imbalance of network traffic, and initiate network optimization, thus helping the network operator or the virtual network service provider to use the network more efficiently and save Capital Expense (CAPEX) and Operating Expense (OPEX).
Customer services have various Service Level Agreement (SLA) requirements, such as service availability, latency, jitter, packet loss rate, Bit Error Rate (BER), etc. The TE network can satisfy service availability and BER requirements by providing different protection and restoration mechanisms. However, for other SLA requirements (like latency), there are no such mechanisms. In order to provide high quality services according to the customer SLA, one possible solution is to measure the SLA related performance parameters, and dynamically provision and optimize services based on the performance monitoring results.
Performance monitoring in a large scale network could generate a huge amount of performance information. Therefore, the appropriate way to deliver the information at the client and network interfaces should be carefully considered.

Design of the Data Models This document describes two YANG models:

(i): TE Telemetry Model which provides the TE-Tunnel level of performance monitoring mechanism and scaling intent mechanism that allows scale in/out programming by the customer. (See & for details).
(ii): VN Telemetry Model which provides the VN level of the aggregated performance monitoring mechanism and scaling intent mechanism that allows scale in/out programming by the customer (See & for details).

TE Telemetry Model This model describes the performance telemetry for the TE tunnel. The telemetry data is augmented to the TE tunnel. This model also allows autonomic traffic engineering scaling intent configuration mechanism on the TE-tunnel level. Various conditions can be set for auto-scaling based on the telemetry data (See for details) As shown in , the TE Telemetry Model augments the TE-Tunnel Model to enhance TE performance monitoring capability. This monitoring capability will facilitate re-optimization and reconfiguration of TE tunnels based on the performance monitoring data collected via the TE Telemetry YANG model.

TE Telemetry Model Relationship

VN Telemetry Model As shown in , the VN Telemetry Model augments the basic VN model to enhance VN monitoring capability. This monitoring capability will facilitate re-optimization and reconfiguration of VNs based on the performance monitoring data collected via the VN Telemetry YANG model. This model also imports TE telemetry model to reuse the groupings.

VN Telemetry Model Relationships This model describes the performance telemetry for the VN model. The telemetry data is augmented to the VN model at the VN Level as well as at the individual VN member level. This model also allows autonomic traffic engineering scaling intent configuration mechanism on the VN level. Scale in/out criteria might be used for network autonomics in order for the controller to react to a certain set of variations in monitored parameters (See for illustrations). Moreover, this model also provides a mechanism to define aggregated VN telemetry parameters as a grouping of underlying VN-member level telemetry parameters. This is unique to the VN model as a VN is made up of multiple VN-members and further each VN-member could be set across multiple TE tunnels. Grouping operation (such as maximum, mean) could be set at the time of configuration. For example, if "maximum" grouping operation is used for delay at the VN level, the VN telemetry data is reported as the maximum of {delay_vn_member_1, delay_vn_member_2,.. delay_vn_member_N}. Thus, this telemetry aggregation mechanism allows the aggregation (or grouping) of a certain common set of telemetry values under a grouping operation. This can also be done at the VN-member level to suggest how the end-to-end (E2E) telemetry be inferred from the per domain tunnels created and monitored by PNCs. The provides an example interaction.

TE Telemetry Model Interactions

VPN Service Performance Monitoring The YANG model in provides network performance monitoring (PM) and VPN service performance monitoring that can be used to monitor and manage network performance on the topology at higher layer or the service topology between VPN sites. Thus the YANG models in this document could be used along side with ietf-network-vpn-pm to understand and correlate the performance monitoring at the VPN service and the underlying TE level.

Autonomic Scaling Intent Mechanism The scaling intent configuration mechanism allows the client to configure automatic scale-in and scale-out mechanisms on both the TE-tunnel and the VN level. Various conditions can be set for auto-scaling based on the PM telemetry data. There are a number of parameters involved in the mechanism:

scale-out-intent or scale-in-intent: whether to scale-out or scale-in.
performance-type: performance metric type (e.g., one-way-delay, one-way-delay-min, one-way-delay-max, two-way-delay, two-way- delay-min, two-way-delay-max, utilized bandwidth, etc.)
threshold-value: the threshold value for a certain performance-type that triggers scale-in or scale-out.
scaling-operation-type: in case where scaling condition can be set with one or more performance types, then scaling-operation-type (AND, OR, MIN, MAX, etc.) is applied to these selected performance types and its threshold values.
Threshold-time: the duration for which the criteria needs to hold true.
Cooldown-time: the duration after a scaling action has been triggered, for which there will be no further operation.

The tree in is a part of ietf-te-telemetry tree whose model is presented in full detail in Sections 6 & 7.

The scaling intent Let's say the client wants to set the scaling out operation based on two performance-types (e.g., two-way-delay and utilized-bandwidth for a te-tunnel), it can be done as follows:

Set Threshold-time: x (sec) (duration for which the criteria must hold true)
Set Cooldown-time: y (sec) (the duration after a scaling action has been triggered, for which there will be no further operation)
Set AND for the scale-out-operation-type

In the scaling condition's list, the following two components can be set: List 1: Scaling Condition for Two-way-delay

performance type: Two-way-delay
threshold-value: z milli-seconds

List 2: Scaling Condition for Utilized bandwidth

performance type: Utilized bandwidth
threshold-value: w megabytes

Refer for some examples of scaling intent.

Performance Monitoring Parameters This model augments Tunnel model to include performance parameters from the grouping performance-metrics-attributes from te-types :

one-way-delay
one-way-delay-normality
one-way-residual-bandwidth
one-way-residual-bandwidth-normality
one-way-available-bandwidth
one-way-available-bandwidth-normality
one-way-utilized-bandwidth
one-way-utilized-bandwidth-normality
two-way-delay
two-way-delay-normality

Performance Monitoring Parameters

Notification This model does not define specific notifications. To enable notifications, the mechanism defined in and can be used. This mechanism currently allows the user to:

Subscribe to notifications on a per client basis.
Specify subtree filters or xpath filters so that only interested contents will be sent.
Specify either periodic or on-demand notifications.

YANG Push Subscription Examples allows subscriber applications to request a continuous, customized stream of updates from a YANG datastore. The example in shows the way for a client to subscribe to the telemetry information for a particular tunnel (Tunnel1). The telemetry parameter that the client is interested in is one-way- delay.

TE Tunnel Subscription Example Tunnel1

500 encode-xml ]]> The example in shows the way for a client to subscribe to the telemetry information for all VNs. The telemetry parameter that the client is interested in is one-way-delay and one-way-utilized- bandwidth.

VN Subscription Example

500 ]]>

Scaling Examples The example in shows the way to configure a TE tunnel with the scaling-out intent to re-optimize when the the scaling condition of two-way-delay crossing 100 milliseconds (100000 microseconds) for a threshold of 1 min (60 seconds).

TE Tunnel Scaling Example Tunnel1

two-way-delay

100000

]]> The example in shows the way to configure a VN with the scaling-in intent to reduce bandwidth when the the scaling condition of utilized-percentage crossing 50 percent for a threshold of 5 minutes (300 seconds).

VN Scaling Example

VN1

300

utilized-percentage

]]> The example in shows the way to configure a VN with the scaling-in when the the scaling condition of one-way-delay-variation crossing 100 milliseconds (100000 microseconds) OR one-way-delay crossing 50 milliseconds (50000 microseconds) for a threshold of 2 minutes (120 seconds).

VN Scaling Example with OR condition

VN2

120

one-way-delay-variation

100000

one-way-delay

50000

]]> The example in shows the way to configure a grouping operation at the VN level to require that the VN level one-way-delay needs to be the reported as the max of the one-way-delay at the VN-member level, where as the utilized-percentage is the mean.

VN Grouping Operation Example

VN1

one-way-delay

maximum

utilized-percentage

mean

]]>

YANG Data Tree

ietf-te-telemetry YANG model tree

ietf-vn-telemetry YANG model tree /te:te/tunnels/tunnel/name +--rw operation* [performance-type] +--rw performance-type identityref +--rw grouping-operation? identityref ]]>

YANG Data Model

ietf-te-telemetry model The YANG code is as follows: WG List: Editor: Young Lee Dhruv Dhody "; description "This module describes YANG data model for performance monitoring parameters (telemetry data) for TE tunnels. Copyright (c) 2023 IETF Trust and the persons identified as authors of the code. All rights reserved. Redistribution and use in source and binary forms, with or without modification, is permitted pursuant to, and subject to the license terms contained in, the Revised BSD License set forth in Section 4.c of the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info). This version of this YANG module is part of RFC XXXX; see the RFC itself for full legal notices."; /* Note: The RFC Editor will replace XXXX with the number assigned to the RFC once draft-ietf-teas-pm-telemetry- autonomics becomes an RFC.*/ revision 2023-03-10 { description "Initial revision."; reference "RFC XXXX: YANG models for VN/TE Performance Monitoring Telemetry and Scaling Intent Autonomics"; } identity telemetry-param-type { description "Base identity for telemetry param types"; } identity one-way-delay { base telemetry-param-type; description "To specify average Delay in one (forward) direction in microseconds. At the VN level, it is the max delay of the VN-members. The threshold-value for this type is interpreted as microseconds."; reference "RFC 7471: OSPF Traffic Engineering (TE) Metric Extensions. RFC 8570: IS-IS Traffic Engineering (TE) Metric Extensions. RFC 7823: Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions"; } identity two-way-delay { base telemetry-param-type; description "To specify average Delay in both (forward and reverse) directions in microseconds. At the VN level, it is the max delay of the VN-members. The threshold-value for this type is interpreted as microseconds."; reference "RFC 7471: OSPF Traffic Engineering (TE) Metric Extensions. RFC 8570: IS-IS Traffic Engineering (TE) Metric Extensions. RFC 7823: Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions"; } identity one-way-delay-variation { base telemetry-param-type; description "To specify average Delay Variation in one (forward) direction in microseconds. At the VN level, it is the max delay variation of the VN-members. The threshold-value for this type is interpreted as microseconds."; reference "RFC 7471: OSPF Traffic Engineering (TE) Metric Extensions. RFC 8570: IS-IS Traffic Engineering (TE) Metric Extensions. RFC 7823: Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions"; } identity two-way-delay-variation { base telemetry-param-type; description "To specify average Delay Variation in both (forward and reverse) directions in microseconds. At the VN level, it is the max delay variation of the VN-members. The threshold-value for this type is interpreted as microseconds."; reference "RFC 7471: OSPF Traffic Engineering (TE) Metric Extensions. RFC 8570: IS-IS Traffic Engineering (TE) Metric Extensions. RFC 7823: Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions"; } identity utilized-bandwidth { base telemetry-param-type; description "To specify utilized bandwidth over the specified source and destination in bytes per second. The threshold-value for this type is interpreted as bytes per second."; reference "RFC 7471: OSPF Traffic Engineering (TE) Metric Extensions. RFC 8570: IS-IS Traffic Engineering (TE) Metric Extensions. RFC 7823: Performance-Based Path Selection for Explicitly Routed Label Switched Paths (LSPs) Using TE Metric Extensions"; } identity utilized-percentage { base telemetry-param-type; description "To specify utilization percentage of the entity (e.g., tunnel, link, etc.)"; } /* Typedef */ typedef scale-op { type enumeration { enum UP { description "Scale up the bandwidth capacity"; } enum DOWN { description "Scale down the bandwidth capacity"; } } description "Scaling operation"; } typedef scaling-criteria-operation { type enumeration { enum AND { description "AND operation"; } enum OR { description "OR operation"; } } description "Operations to analyze list of scaling criteria"; } typedef scale-value { type union { type uint32; type rt-types:bandwidth-ieee-float32; type rt-types:percentage; type te-types:te-bandwidth; } description "Union of scale values of various types"; } grouping scaling-duration { description "Base scaling criteria durations"; leaf threshold-time { type uint32; units "seconds"; description "The duration for which the criteria must hold true. The value of '0' indicates an immediate scaling with no duration to wait."; } leaf cooldown-time { type uint32; units "seconds"; description "The duration after a scaling-in/scaling-out action has been triggered, for which there will be no further operation. The value of '0' indicates an immediate scaling action with no duration to wait."; } } grouping scaling-criteria { description "Grouping for scaling criteria"; leaf performance-type { type identityref { base telemetry-param-type; } description "Reference to the tunnel level telemetry type"; } leaf threshold-value { type scale-value; description "Scaling threshold for the telemetry parameter type. The value is it be interpreted as per the type."; } } grouping scaling-in-intent { description "Basic scaling in intent"; uses scaling-duration; list scaling-condition { key "performance-type"; description "Scaling conditions"; uses scaling-criteria; leaf scale-in-operation-type { type scaling-criteria-operation; default "AND"; description "Operation to be applied to check between scaling criteria to check if the scale in threshold condition has been met. Defaults to AND"; } } leaf scale-in-op { type scale-op; default "DOWN"; description "The scaling operation to be performed when scaling condition is met"; } leaf scale { type scale-value; description "Additional scaling-by information to be interpreted as per the scale-in-op."; } } grouping scaling-out-intent { description "Basic scaling out intent"; uses scaling-duration; list scaling-condition { key "performance-type"; description "Scaling conditions"; uses scaling-criteria; leaf scale-out-operation-type { type scaling-criteria-operation; default "OR"; description "Operation to be applied to check between scaling criteria to check if the scale out threshold condition has been met. Defauls to OR"; } } leaf scale-out-op { type scale-op; default "UP"; description "The scaling operation to be performed when scaling condition is met"; } leaf scale { type scale-value; description "Additional scaling-by information to be interpreted as per the scale-out-op."; } } augment "/te:te/te:tunnels/te:tunnel" { description "Augmentation parameters for config scaling-criteria TE tunnel topologies. Scale in/out criteria might be used for network autonomics in order the controller to react to a certain set of monitored params."; container te-scaling-intent { description "The scaling intent"; container scale-in-intent { description "scale-in"; uses scaling-in-intent; } container scale-out-intent { description "scale-out"; uses scaling-out-intent; } } container te-telemetry { config false; description "Telemetry Data"; uses te-types:performance-metrics-attributes; } } } ]]>

ietf-vn-telemetry model The YANG code is as follows: WG List: Editor: Young Lee Dhruv Dhody "; description "This module describes YANG data models for performance monitoring parameters (telemetry data) for Virtual Network (VN). Copyright (c) 2023 IETF Trust and the persons identified as authors of the code. All rights reserved. Redistribution and use in source and binary forms, with or without modification, is permitted pursuant to, and subject to the license terms contained in, the Revised BSD License set forth in Section 4.c of the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info). This version of this YANG module is part of RFC XXXX; see the RFC itself for full legal notices."; /* Note: The RFC Editor will replace XXXX with the number assigned to the RFC once draft-lee-teas-pm-telemetry- autonomics becomes an RFC.*/ revision 2023-03-10 { description "Initial revision."; reference "RFC XXXX: YANG models for VN/TE Performance Monitoring Telemetry and Scaling Intent Autonomics"; } identity grouping-op { description "Base identity for grouping-operation"; } identity minimum { base grouping-op; description "Select the minimum of the monitored parameters"; } identity maximum { base grouping-op; description "The maximum of the monitored parameters"; } identity mean { base grouping-op; description "The mean of the monitored parameters"; } identity standard-deviation { base grouping-op; description "The standard deviation of the monitored parameters"; } identity sum { base grouping-op; description "The sum of the monitored parameters"; } identity and { base grouping-op; description "Logical AND operation"; } identity or { base grouping-op; description "Logical OR operation"; } grouping grouping-operation { list operation { key "performance-type"; leaf performance-type { type identityref { base te-tel:telemetry-param-type; } description "Reference to the tunnel level telemetry type"; } leaf grouping-operation { type identityref { base grouping-op; } description "describes the operation to apply to the underlying TE tunnels"; } description "Grouping operation for each performance-type"; } description "Grouping operation for each performance-type"; } augment "/vn:virtual-network/vn:vn" { description "Augmentation parameters for state TE VN topologies."; container vn-scaling-intent { description "scaling intent"; container scale-in-intent { description "VN scale-in"; uses te-tel:scaling-in-intent; } container scale-out-intent { description "VN scale-out"; uses te-tel:scaling-out-intent; } } container vn-telemetry { description "VN telemetry params"; container params { config false; description "Read-only telemetry parameters"; uses te-types:performance-metrics-attributes; } uses grouping-operation; } } augment "/vn:virtual-network/vn:vn/vn:vn-member" { description "Augmentation parameters for state TE vn member topologies."; container vn-member-telemetry { description "VN member telemetry params"; container params { config false; description "Read-only telemetry parameters"; uses te-types:performance-metrics-attributes; leaf-list te-tunnel-ref { type leafref { path "/te:te/te:tunnels/te:tunnel/te:name"; } description "A list of underlying TE tunnels that form the VN-member"; } } uses grouping-operation; } } } ]]>

Security Considerations The YANG modules specified in this document defines a schema for data that is designed to be accessed via network management protocols such as NETCONF or RESTCONF . The lowest NETCONF layer is the secure transport layer, and the mandatory-to-implement secure transport is Secure Shell (SSH) . The lowest RESTCONF layer is HTTPS, and the mandatory-to-implement secure transport is TLS . The Network Configuration Access Control Model (NACM) provides the means to restrict access for particular NETCONF or RESTCONF users to a preconfigured subset of all available NETCONF or RESTCONF protocol operations and content. There are a number of data nodes defined in this YANG module that are writable/creatable/deletable (i.e., config true, which is the default). These data nodes may be considered sensitive or vulnerable in some network environments. Write operations (e.g., edit-config) to these data nodes without proper protection can have a negative effect on network operations. These are the subtrees with the write operation that can be exploited to impact the network monitoring. An incorrect condition could cause frequent scaling operation to be executed causing harm to the network:

/te:te/te:tunnels/te:tunnel/te-scaling-intent/scale-in-intent
/te:te/te:tunnels/te:tunnel/te-scaling-intent/scale-out-intent
/vn:virtual-network/vn:vn/vn-scaling-intent/scale-in-intent
/vn:virtual-network/vn:vn/vn-scaling-intent/scale-out-intent

Further, following are the subtrees with the write operation that can be exploited by setting an incorrect grouping operation for the VN operation impacting the network monitoring:

/vn:virtual-network/vn:vn/vn-telemetry/operation
/vn:virtual-network/vn:vn/vn:vn-member/vn-member-telemetry/operation

Some of the readable data nodes in this YANG module may be considered sensitive or vulnerable in some network environments. It is thus important to control read access (e.g., via get, get-config, or notification) to these data nodes. These are the subtrees with the read operations that can be exploited to learn real-time (and sensitive) telemetry information about the TE tunnels and VN:

/te:te/te:tunnels/te:tunnel/te-telemetry
/vn:virtual-network/vn:vn/vn-telemetry
/vn:virtual-network/vn:vn/vn:vn-member/vn-member-telemetry

IANA Considerations This document registers the following namespace URIs in the IETF XML registry : This document registers the following YANG modules in the YANG Module Names registry :

Acknowledgments We thank Adrian Farrel, Rakesh Gandhi, Tarek Saad, Igor Bryskin, Kenichi Ogaki, and Greg Mirsky for useful discussions and their suggestions for this work. Thanks to Reshad Rahman for an excellent YANGDOCTOR review.

References Normative References Informative References

Out of Scope This document exclusively focus on performance monitoring telemetry and scaling intent mechanisms of the underlying transport (TE-tunnels and Virtual Networks (VNs)). The performance monitoring of the services is out of scope. See for details about VPN performance monitoring. Similarly performance monitoring of IETF network slices could be developed and it is clearly out of scope of this document.

Contributors The following have contributed significantly and should be considered as co-author: