YANG Modules for Service AssuranceCisco Systems, Inc.De Kleetlaan 6a b11831 DiegemBelgiumbclaise@cisco.comCisco Systems, Inc.1, rue Camille Desmoulins92782 Issy Les MoulineauxFrancejquilbeu@cisco.comNTTSiriusdreef 70-72HoofddorpWT2132Netherlandspaolo@ntt.netTIM S.p.Avia G. Reiss Romoli, 27410148 TorinoItalypaolo2.fasano@telecomitalia.it
OPS
OPSAWG
This document proposes YANG modules for the Service Assurance for Intent-based Networking Architecture.
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
when, and only when, they appear in all capitals, as shown here.
The terms used in this document are defined in draft-claise-opsawg-service-assurance-architecture IETF draft.
The "Service Assurance for Intent-based Networking Architecture" draft-claise-opsawg-service-assurance-architecture, specifies the framework and all of its components for service assurance. This document complements the architecture by providing open interfaces between components. More specifically, the goal is to provide YANG modules for the purpose of service assurance in a format that is:
machine readablevendor independentaugmentable
The main YANG module, ietf-service-assurance, defines objects for assuring network services based on their decomposition into so-called subservices. The subservices are hierarchically organised by dependencies. The subservices, along with the dependencies, constitute an assurance graph. This module should be supported by an agent, able to interact with the devices in order to produce a health status and symptoms for each subservice in the assurance graph.
This module is intended for the following use cases:
Assurance graph configuration:
Subservices: configure a set of subservices to assure, by specifying their types and parameters.
Dependencies: configure the dependencies between the subservices, along with their type.
Assurance telemetry: export the health status of the subservices, along with the observed symptoms.
The second YANG module, ietf-service-assurance-device, extends the ietf-service-assurance module to add support for the subservice DeviceHealthy. Additional subservice types might be added the same way.
The third YANG module, example-service-assurance-device-acme, extends the ietf-service-assurance-device module as an example to add support for the subservice DeviceHealthy, with specifics for the fictional ACME Corporation. Additional vendor-specific parameters might be added the same way.
The following tree diagram
provides an overview of the ietf-service-assurance data model.
The ietf-service-assurance YANG model assumes an identified number of subservices, to be assured independently. A subservice is a feature or a subpart of the network system that a given service instance might depend on. Example of subservices include:
DeviceHealthy: whether a device is healthy, and if not, what are the symptoms. Potential symptoms are "CPU overloaded", "Out of RAM", or "Out of TCAM".
ConnectivityHealthy: given two IP addresses owned by two devices, what is the quality of the connection between them. Potential symptoms are "No route available" or "ECMP Imbalance".
The first example is a subservice representing a subpart of the network system, while the second is a subservice representing a feature of the network, In both cases, these subservices might depend on other subservices, for instance, the connectivity might depend on a subservice representing the routing mechanism and on a subservice representing ECMP.
The symptoms are listed for each subservice. Each symptom is specified by a unique id and contains a health-score-weight (the impact to the health score incurred by this symptom), a label (text describing what the symptom is), and dates and times at which the symptom was detected and stopped being detected. While the unique id is sufficient as an unique key list, the start-date-time second key help sorting and retrieving relevant symptoms.
The assurance of a given service instance can be obtained by composing the assurance of the subservices that it depends on, via the dependency relations.
In order to declare a subservice MUST provide:
A type: identity inheriting of the base identity for subservice,
An id: string uniquely identifying the subservice among those with the same identity,
Some parameters, which should be specified in an augmenting model, as described in the next sections.
The type and id uniquely identify a given subservice. They are used to indicate the dependencies. Dependencies have types as well. Two types are specified in the model:
Impacting: such a dependency indicates an impact on the health of the dependent,Informational: such a dependency might explain why the dependent has issues but does not impact its health.
To illustrate the difference between "impacting" and "informational", consider the subservice InterfaceHealthy, representing a network interface. If the device to which the network interface belongs goes down, the network interface will transition to a down state as well. Therefore, the dependency of InterfaceHealthy towards DeviceHealthy is "impacting". On the other hand, as a the dependency towards the ECMPLoad subservice, which checks that the load between ECMP remains ce remains stable throughout time, is only "informational". Indeed, services might be perfectly healthy even if the load distribution between ECMP changed. However, such an instability might be a relevant symptom for diagnosing the root cause of a problem.
Service instances MUST be modeled as a particular type of subservice with two parameters, a type and an instance name. The type is the name of the service defined in the network orchestrator, for instance "point-to-point-l2vpn". The instance name is the name assigned to the particular instance that we are assuring, for instance the name of the customer using that instance.
The "under-maintenance" and "maintenance-contact" flags inhibit the emission of symptoms for that subservice and subservices that depend on them. See Section 3.7 of for a more detailed discussion.
By specifying service instances and their dependencies in terms of subservices, one defines the whole assurance to apply for them. An assurance agent supporting this model should then produce telemetry in return with, for each subservice: a health-status indicating how healthy the subservice is and when the subservice is not healthy, a list of symptoms explaining why the subservice is not healthy.
<CODE BEGINS> file "ietf-service-assurance@2020-01-13.yang"<CODE ENDS>
The following tree diagram
provides an overview of the ietf-service-assurance-device data model.
The following tree diagram
provides an overview of the ietf-service-assurance and ietf-service-assurance-device data models.
As the number of subservices will grow over time, the YANG module is designed to be extensible. A new subservice type requires the precise specifications of its type and expected parameters. Let us illustrate the example of the new DeviceHealthy subservice type. As the name implies, it monitors and reports the device health, along with some symptoms in case of degradation.
For our DeviceHealthy subservice definition, the new device-idty is specified, as an inheritance from the base identity for subservices. This indicates to the assurance agent that we are now assuring the health of a device.
The typical parameter for the configuration of the DeviceHealthy subservice is the name of the device that we want to assure. By augmenting the parameter choice from ietf-service-assurance YANG module for the case of the device-idty subservice type, this new parameter is specified.
<CODE BEGINS> file "ietf-service-assurance-device@2020-01-13.yang"<CODE ENDS>
The following tree diagram
provides an overview of the ietf-service-assurance-interface data model.
The following tree diagram
provides an overview of the ietf-service-assurance, ietf-service-assurance-device, and
ietf-service-assurance-interface data models.
For our InterafaceHealthy subservice definition, the new interface-idty is specified, as an inheritance from the base identity for subservices. This indicates to the assurance agent that we are now assuring the health of an interface.
The typical parameters for the configuration of the InterfaceHealthy subservice are the name of the device and, on that specific device, a specific interface. By augmenting the parameter choice from ietf-service-assurance YANG module for the case of the interface-idty subservice type, those two new parameter are specified.
<CODE BEGINS> file "ietf-service-assurance-interface@2020-01-13.yang"<CODE ENDS>
The following tree diagram
provides an overview of the example-service-assurance-device-acme data model.
The following tree diagram
provides an overview of the ietf-service-assurance, ietf-service-assurance-device, and example-service-assurance-device-acme data models.
Under some circumstances, vendor-specific subservice types might be required. As an example of this vendor-specific implementation, this section shows how to augment the ietf-service-assurance-device module to add support for the subservice DeviceHealthy, specific to the ACME Corporation. The new parameter is acme-specific-parameter.
The YANG module 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 and data nodes and their sensitivity/vulnerability:
/subservices/subservice/type/subservices/subservice/id/subservices/subservice/under-maintenance/subservices/subservice/maintenance-contactThis document registers two URIs in the IETF XML
registry . Following the format in
, the following registrations are
requested:This document registers three YANG modules in the
YANG Module Names registry .
Following the format in , the
the following registrations are requested:-Nonedraft-claise-opsawg-service-assurance-architectureCisco SystemsCisco Systemsv04 - v05
Added the concept of symptoms-history-startChanged label to description, under symptoms. This was
confusing as there was two labels in the modelsv03 - v04
Add the interface subservice, with two parametersv02 - v03
Added the maintenace window conceptsv01 - v02
Improved leaf namingClarified some concepts: symptoms, dependencyv00 - v01
Terminology clarificationsProvide example of impacting versus impacted dependenciesThe authors would like to thank Jan Lindblad for his help during the design of these YANG modules. The authors would like to thank Stephane Litkowski and Charles Eckel for their reviews.