New component capabilities for Frankfurt, i.e. the functional enhancements.
Frankfurt release targets the completion of the work started in Dublin to re-build the Policy Framework infrastructure. The new capabilities were to complete missing functionality that was previously in the legacy components, or was targeted to be part of the platform and was not able to be finished in Dublin:
Consolidation of Policy Health Check - China Ericsson Team
POLICY-2025 - Getting issue details... STATUS - Per Integration team, one single health check requested that gives status for the whole of Policy Framework components.
- New API implemented in PAP : /policy/pap/v1/pdps/healthcheck
Example API result:
{
"pdps": {
"xacml": [
{
"instanceId": "CN-00000994",
"pdpState": "ACTIVE",
"healthy": "HEALTHY"
}
],
"drools": [
{
"instanceId": "CN-00000994",
"pdpState": "ACTIVE",
"healthy": "HEALTHY"
}
],
"apex": [
{
"instanceId": "CN-00000994",
"pdpState": "ACTIVE",
"healthy": "HEALTHY"
}
]
},
"healthy": true,
"api": {
"name": "Policy API",
"url": "http://policy-api:6969/policy/api/v1/healthcheck",
"healthy": true,
"code": 200,
"message": "alive"
},
"distribution": {
"name": "Policy SSD",
"url": "http://policy-distribution:6969/healthcheck",
"healthy": true,
"code": 200,
"message": "alive"
},
"pap": {
"name": "Policy PAP",
"url": "http://policy-pap:6969/policy/pap/v1/healthcheck",
"healthy": true,
"code": 200,
"message": "alive"
}
}
PDP Statistics - China Ericsson Team
POLICY-2025 - Getting issue details... STATUS - The same epic also covers the work for collecting various statistics of registered PDP engines and storing them in the Policy DB.
- Policy models and related provider classes created for storing PDP Statistics in DB.
- REST API created in PAP to fetch the PDP Statistics saved in Policy DB.
- PdpStatus heartbeat messages from apex-pdp has been updated to include current statistics in every interval. Xacml-pdp & drools-pdp will be updated shortly.
- PAP handlers has been updated to collect & store the statistics coming in heartbeat messages to Policy DB.
PDP Monitoring GUI - China Ericsson Team
POLICY-1751 - Getting issue details... STATUS - This epic covers the work to create a GUI that polls the PDP statistics from PAP for monitoring PDP engines. This is currently work in progress.
A new UI is being developed to show
- the list of PDP engines registered with PAP
- the statistics of an individual PDP engine.
- the cumulative statistics of a group of PDP engines.
- the cumulative statistics of all PDP engines for a particular domain.
Integration of CDS As an Actor Former user (Deleted) Rashmi Pujar
POLICY-2087 - Getting issue details... STATUS - Started in Dublin as dark code, finishing the support for this new actor to be fully available for Control Loop Operational Policies
Policy Update Notifications James Hahn
POLICY-1840 - Getting issue details... STATUS - When policies are deployed/undeployed (eg new version or removal), then some ONAP components need more immediate notification. Notably, DCAE microservices need to respond quickly to policy changes. James Hahn
Policy Validation Liam Fallon
POLICY-2027 - Getting issue details... STATUS - Enhances the validation of TOSCA Policy Types and provides tools to Policy Designers. Liam Fallon
Better support for TOSCA Data Types, Policy Types and Policies. Currently (Dublin/El Alto), policies are validated by PDPs at load time, and the Policy Framework checks that the policy type of a policy is valid and exists when it is stored over the API
In Frankfurt, we will:
- Validate that the parent policy types, and referenced data types in policy types and data types exist
- Return fully dereferenced policy types and data types on get requests
- Vaidate consistent create and update on policy types and data types
- Validate references to policy tpyes and data types prior to deletion of policy types and data types
- Validate versions of policy types and policies
- Validation of policies against their policy types on create and update of policies (Stretch)
See R6 Data Type, Policy Type, and Policy Referencing and Versioning for more information.
Support for Native Policies Chenfei Gao Ram Krishna Verma Pamela Dragosh
POLICY-1845 - Getting issue details... STATUS - Some support for this in the legacy components, this work greatly enhances the ability to load PDP native policies that either execute alone or in conjunction with TOSCA Policy Types. Adds in Apex Native Policy Types in addition to XACML, Drools.
- Designed TOSCA-compliant solution to create/read/update/delete(CRUD), deploy/undeploy and store native PDP policies and their artifacts
- Designed native PDP policy types off which native policies executable in different PDP engines (i.e. Drools, XACML, APEX) can be created and deployed
- Build a validator for the API payload provided to create native PDP policies to avoid unnecessary inconsistency between policy artifacts in the nexus and TOSCA policies in the database
- Build Drools PDP (PDP-D) support for native Drools policy execution
- Build XACML PDP (PDP-X) support for native XACML policy enforcement
- Build APEX PDP (PDP-A) support for native APEX policy execution
Optimization and Naming Policies - Better Use of the Decision API Pamela Dragosh REQ-162 - Getting issue details... STATUS
In working with the OOF team, we are moving some of the decision making that the legacy engine could not support into the new xacml PDP engine which supports a Decision API. Previously, OOF had to make 2 separate API calls.
The addition of "matchable" attributes in the Optimization and Naming policies allow the XACML PDP to dynamically translate TOSCA policies to XACML for more fine-grained policy Decisions.
The OOF optimize uses a "best matches" algorithm on top of the XACML engine fine-grained decision in order to further refine the decision to return only the "best matching" policy given the attributes in the Decision API call.
tosca_definitions_version: tosca_simple_yaml_1_0_0
policy_types:
onap.policies.Optimization:
derived_from: tosca.policies.Root
version: 1.0.0
description: The base policy type for all policies that govern optimization
properties:
scope:
description: Scope for the policy - could be for a specific release.
type: list
metadata:
matchable: true
required: true
entry_schema:
type: string
geography:
description: One or more geographic regions
type: list
metadata:
matchable: true
required: true
entry_schema:
type: string
constraints:
- valid_values:
- US
- International
identity:
description: Used internally for identification
type: string
required: true
The OOF optimization Decision includes the ability to ask the decision to take in "context" when returning policy decisions. Eliminates one less call the OOF project had to make to the Policy Decision API to retrieve policies during runtime.
{
"ONAPName": "OOF",
"ONAPComponent": "OOF-component",
"ONAPInstance": "OOF-component-instance",
"context" : {
"subscriberName": []
},
"action": "optimize",
"resource": {
"scope": [],
"services": [],
"resources": [],
"geography": []
}
}
TOSCA Compliant Operational and Guard Policies Pamela Dragosh REQ-21 - Getting issue details... STATUS
Finishing the work started in Dublin to fully move ALL the Control Loop Policy Types into TOSCA compliance with CLAMP team. Previously our operational and guard policies used a YAML format defined a few years ago. In this effort we are translating it to be fully compliant with TOSCA and cleaning up some of the syntax.
tosca_definitions_version: tosca_simple_yaml_1_0_0
policy_types:
onap.policies.controlloop.operational.Common:
derived_from: tosca.policies.Root
version: 1.0.0
description: Operational Policy for Control Loop execution
properties:
id:
type: String
description: The unique control loop id.
required: true
timeout:
type: Integer
description: |
Overall timeout for executing all the operations. This timeout should equal or exceed the total
timeout for each operation listed.
required: true
abatement:
type: Boolean
description: Whether an abatement event message will be expected for the control loop from DCAE.
required: true
default: false
trigger:
type: String
description: Initial operation to execute upon receiving an Onset event message for the Control Loop.
required: true
operations:
type: List
description: List of operations to be performed when Control Loop is triggered.
required: true
entry_schema:
type: onap.datatype.controlloop.Operation
onap.policies.controlloop.operational.common.Apex:
derived_from: onap.policies.controlloop.operational.Common
type_version: 1.0.0
version: 1.0.0
description: Operational policies for Apex PDP
properties:
onap.policies.controlloop.operational.common.Drools:
derived_from: onap.policies.controlloop.operational.Common
type_version: 1.0.0
version: 1.0.0
description: Operational policies for Drools PDP
properties:
controllerName:
type: String
description: Drools controller properties
required: false
data_types:
# TBD if this is needed
onap.datatype.controlloop.operation.Failure:
derived_from: tosca.datatypes.Root
description: Captures information of an operational failure performed for control loop
properties:
messages:
type: String
description: error message
required: true
category:
type: String
description: |
The category the error occurred in. Whether this is a general error from the actor, or the operation
timed out, retries were exhausted in trying to execute the operation, a guard policy prevented the
operation from occuring, or an exception in the system caused the failure.
constraints:
- valid_values: [error, timeout, retries, guard, exception]
onap.datatype.controlloop.Target:
derived_from: tosca.datatypes.Root
description: Definition for a entity in A&AI to perform a control loop operation on
properties:
targetType:
type: String
description: Category for the target type
required: true
constraints:
- valid_values: [VNF, VM, VFMODULE, PNF]
entityIds:
type: Map
description: |
Map of values that identify the resource. If none are provided, it is assumed that the
entity that generated the ONSET event will be the target.
required: false
onap.datatype.controlloop.Actor:
derived_from: tosca.datatypes.Root
description: An actor/operation/target definition
properties:
actor:
type: String
description: The actor performing the operation.
required: true
operation:
type: String
description: The operation the actor is performing.
required: true
target:
type: String
description: The resource the operation should be performed on.
required: true
metadata:
clamp_possible_values: <string:see clamp project for syntax>
payload:
type: Map
description: Name/value pairs of payload information passed by Policy to the actor
required: false
entry_schema:
type: String
onap.datatype.controlloop.Operation:
derived_from: tosca.datatypes.Root
description: An operation supported by an actor
properties:
id:
type: String
description: Unique identifier for the operation
required: true
description:
type: String
description: A user-friendly description of the intent for the operation
required: false
operation:
type: onap.datatype.controlloop.Actor
description: The definition of the operation to be performed.
required: true
metadata:
clamp_possible_values: <string:see clamp project for syntax>
timeout:
type: Integer
description: The amount of time for the actor to perform the operation.
required: true
retries:
type: Integer
description: The number of retries the actor should attempt to perform the operation.
required: true
default: 0
success:
type: String
description: Points to the operation to invoke on success. A value of "final_success" indicates and end to the operation.
required: false
default: final_success
failure:
type: String
description: Points to the operation to invoke on Actor operation failure.
required: false
default: final_failure
failure_timeout:
type: String
description: Points to the operation to invoke when the time out for the operation occurs.
required: false
default: final_failure_timeout
failure_retries:
type: String
description: Points to the operation to invoke when the current operation has exceeded its max retries.
required: false
default: final_failure_retries
failure_exception:
type: String
description: Points to the operation to invoke when the current operation causes an exception.
required: false
default: final_failure_exception
failure_guard:
type: String
description: Points to the operation to invoke when the current operation is blocked due to guard policy enforcement.
required: false
default: final_failure_guard
Much discussion for this was done in the Control Loop sub committee meetings. As captured here: Control Loop Policy Type Design Proposal#1: Simple upgrade of legacy policy types to TOSCA
Policy API Enhancements Chenfei Gao
POLICY-2028 - Getting issue details... STATUS - Finished work such as added "application/yaml" as Content-Type.
- Added the support of "application/yaml" as Content-Type in policy lifecycle API
- Created a new simplified API for creating multiple policies in one shot
- Made policy types preloading configurable in the helm charts
- Added new version validation for POST API payload to constrain clients to carefully specify a reasonable version for policy or policy type
- Preload default policies for ONAP components
PAP ↔ PDP Communication Ram Krishna Verma James Hahn
POLICY-2026 - Getting issue details... STATUS - Clarified PAP ↔ communication of Policy Types, Health Status, PDP Group Details, Passive vs Active modes. Ram Krishna Verma
- Mapping of kubernetes deployment to PDP Subgroup
- Re-synch of PAP and PDP when the supported policy type changes
- Change the PAP Group Deploy/Undeploy API to only create/update PdpGroups
- Create new API to deploy/undeploy policies into PdpGroups
- Handle derivation in PDP supported types