1. High Level Component Definition and Architectural Relationships
The Configuration Persistence Service (CPS) provides storage for real-time run-time configuration and operational parameters that need to be used by ONAP.
In R8, Honolulu, the CPS is a stand-alone component. Project page describing eventual scope and ambition is here: Configuration Persistence Service Project
This page reflects the ambitions for Honolulu.
draw.io Diagram | ||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
2. API definitions
Configuration Persistence Service provides the following interfaces:
Section | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Note: xxxI interface is a Component internal interface. xxxxE interface is a component external interface
CPS consumes the following Interfaces:
Section | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
3. Configuration Persistence Service Component Description:
A more detailed figure and description of the component.
PURPOSE:
REPOSITORY - The types of data that is stored in the Run-Time data storage repository for:
(1) CONFIGURATION PARAMETERS used by xNFs in run time. For example 5G Network run-time instance configuration information. and
(2) OPERATIONAL PARAMETERS used by ONAP and xNFs. Exo-inventory information is information that doesn't belong in A&AI.
- (3) COMPONENT DATA used by ONAP components to persist data does not belong in A&AI and is not part of the xNF models.
- DATA LAKE - It is designed to be a common services data layer which can serve as a data lake for xNF data.
- SYNCING - The RunTime DB enables the ability to sync data between ONAP & the xNFs. (The source of truth can be defined). (FUTURE)
- CM FUNCTIONS - Enables OSS configuration, optimization, and LCM operations. (FUTURE)
- CM FUNCTIONS - Enables future CM & Data management functions such as xNF Crash restoration, data restoration, data history management and auditing. (FUTURE)
- CENTRAL/DISTRIBUTED - Because it is a common service, it is part of an ONAP installation, so it could be deployed with either an Edge ONAP installation or a centralized ONAP installation. (FUTURE)
- SCOPE - The Run Time DB could also serve as the data storage to store for example ONAP Policy Rules, CLAMP Control Loop, Operational Views (FUTURE) and also accommodate other resources.
ACCESS (READ/WRITE):
- xNF Data - Run-Time parameters can be READ/WRITE by any ONAP platform component and any ONAP plug-in via the xNF data access interface. Examples of ONAP platform components are A&AI, SDC, SDNC etc.
Non-xNF Data - Parameters can be READ/WRITE by owner.
- ACCESS CONTROL - The data owner may grant permission to other components to READ and/or WRITE non-xNF data. (FUTURE)
SYNCING (INVENTORY):
ELEMENT SYNC - Software keeps the A&AI elements with the elements in the RunTime DB in Sync.
- A&AI - A&AI is still the master of valid entities in the network and provides a dynamic view of the assets (xNFs) available to ONAP
- CPS (xNF Proxy) - The CPS is the owner (within ONAP) of the associate (exo-inventory) data associated with the xNFs.
- DYNAMIC VIEW - When a xNF appears or is removed from the system, CPS records will be added/removed based on A&AI entries.
3.1 Concepts
A dataspace is a logical separation of data within the CPS. Dataspaces are created by data owners.
An anchor is a logical separation of data within a dataspace. An anchor is associated with a set of YANG modules (model). Models are used to validate data. Anchors are created by data owners.
Anchor and dataspace concepts are described in a presentation here.
An owner is an ONAP component that is responsible for data in the CPS. Responsibility includes deciding how data is exposed to other ONAP components. Options being:
- Data is exposed via µService APIs – no direct access to data via CPS interfaces
- Data is exposed via CPS interfaces – requires registration/discovery and access control to be implemented (FUTURE)
Data ownership and access control are the mechanisms provided by CPS for ONAP to avoid uncontrolled coupling in the data layer.
4. Known system limitations
In Honolulu, the inventory, xNF models, and initial data will be statically defined in files. This will be coordinated with other components participating in the E2E network slicing use case. Changes to these definitions will require a redeployment of CPS.
5. Used Models
RunTime DB uses the following models:
- Inventory Model (Run time platform data model)
- ran-network@2020-08-06.yang – static YANG model for Honolulu
6. System Deployment Architecture
CPS will be deployed as two micro services:
- PostgreSQL µS – requiring a physical volume
- CPS µS – providing all interfaces described above
7. New Capabilities in this Release
This release, CPS adds the following Capabilities:
- Model LCM
- Generic data CRUD
- xNF data CRUD
- xNF data write forwarding
8. References
- The R7 PoC is described here: High Level Design (PoC)
- Ongoing development: Configuration Persistence Service Developer's Landing Page
- CPS backlog: https://jira.onap.org/secure/RapidBoard.jspa?rapidView=228&view=planning.nodetail&quickFilter=713&issueLimit=100
- Decision log: Issues decisions and assumptions