SUPA CHARTER PROPOSAL (with feedback till 2015-08-12, bclaise's temp file, for tracking purposes) Policy is an important configurable component in the delivery of services and the operation of networks. Operators want and need to be able to determine the policies that apply to their different customers and to the equipment that comprises their physical and virtual networks. As policy spans such a wide range of services and device types, it will be helpful if there is a common way of expressing and describing policies that is uniform and consistent in all environments. Such an approach will help to avoid configuration errors that arise from confusion between different systems, will enable easy understanding of policies that apply in different environments, will make the implementation of policy-based systems quicker and cheaper, and will facilitate the rapid development of standards-based data models that include policy elements. SUPA (Simplified Use of Policy Abstractions) defines a model to interface with a network management function (within a controller, an orchestrator, a network element) that takes high-level, possibly network-wide policies as input. SUPA will not define what the function does with that input but anticipates that it will ultimately result in network configuration changes. Practically, SUPA will define base YANG data models to encode policy, which will point to device-, technology-, and service-specific YANG models of other working groups. SUPA will be designed to work with device, protocol, network, and service data models. This working group will be a success when the SUPA policy constructs are re-used in future IETF specifications (and, ideally, specifications from other SDOs), in a manner that will save development time and avoid inconsistencies between data models developed by different working groups. In the meantime, SUPA should not impede work in other working groups while waiting for SUPA to produce its deliverables. The SUPA working group will develop a model for expressing policy at different levels of abstraction. Specifically, three models are envisioned: (i) a generic model that defines concepts and vocabulary needed by policy management independent content of the policy, (ii) a more specific model that refines the generic model to specify how to build policy rules of the event-condition-action paradigm, and (iii) the working group will investigate the practically of building policy rules that declaratively specify what goals to achieve (this is often called "intent-based" policy) but not how to achieve those goals and may create a model accordingly. If the working group finds it necessary to work on an information model before the data model, to help provide guidance and derive the data models, it might do so. The working group will decide later whether the information model needs to be published as an RFC. Out of scope of this working group is: - The specification of a new policy language. The YANG data models are to be used with protocols such as NETCONF/RESTCONF. - Design of protocol-specific policies and specific design for embedded policies in network elements (which are usually interpreted in isolation, and often at timescales that require optimization for specific purposes). - Specific handling of policies (although the application document will provide some examples), and therefore the specification of a policy engine that maps a specific policy instance to actual configuration snippets. List of work items: 1) A document that explains the scope of the policy-based management framework and how it relates to existing work of the IETF. 2) If the working group considers it necessary, an information model composed of policy concepts and vocabulary. 3) A set of YANG data models consisting of a base policy model for representing policy management concepts independent of the type or structure of a policy, plus an extension for defining policy rules according to the event-condition-action paradigm. Another extension for defining policy rules according to a declarative, or intent-based, paradigm, may be produced. These models will be designed to be generic and extensible. 4) An applicability document providing a few examples that demonstrate how the YANG policy data models can be used to express policies that are relevant for network operators. The examples may tie into configuration models or network service models developed by other working groups. The working group will decide how the work items are best mapped into deliverables. The working group will communicate with other SDOs (MEF, TMF, ETSI) that are working on related issues. Milestones: TBD