...
- Tony Finnerty
- Joanne Liu Rudel
- user-4e793 (Bell Canada)Ted Johnson
- Zu Qiang (Ericsson)
- Melanie Sater
- Joanne Liu Rudel
- Michela Bevilacqua
- Mike ElliottMunir Ahmad (Bell Canada)
- --
ARCH WORK
ARCHITECTURE WORK | WIKI LINK |
ARCHITECTURE FLOWS | ARCHCOM: InfoFlow - RunTime Config DB Information Flow |
COMPONENT DESCRIPTION | ARC RunTime DB Component Description - R6 Frankfurt |
PROJECT PROPOSAL | RunTime Config DB Project Proposal (Oct 25 2019) |
DISCUSSION
Topic | Discussion | Calendar Entry | |
---|---|---|---|
R6 CCSDK-based Solution | Project as part of CCSDK ( Yuriy Malakov )
| ||
R7 Project Proposal | RunTime Config DB Project Proposal (Oct 25 2019) Project Proposal work to be done during R6. Project Proposal for R7. Presentation at Arch S/C and TSC during R6.
| ||
R7 Separate Component |
Email from Dan Timoney My understanding from Sandeep was that this work was very much a stretch for Frankfurt. So, I’m okay with work starting in Frankfurt, as long as its structured so that it’s a separable component (i.e. as long as, if it’s not completed in Frankfurt, the platform is not fundamentally broken). I would NOT support creating a separate repository, since there is a fair amount of overhead involved in maintaining each repository on an ongoing basis – both machine and human resources. The Linux Foundation itself has been pushing back on the number of repositories the ONAP projects have and there is now a new approval process needed in order to add new ones. If a new repository is needed, then this team will need to convince me why no existing repository can be used AND will need to provide a resource who is willing to maintain that repository (i.e dealing with security vulnerabilities; policing code coverage ; doing release builds, etc). | ||
R7 Guilin Content / requirements | Requests for R7 Requirements are up. Guilin release - functional requirements proposed list Timeline - Sign-off for R6 is May 7. Historically M0 kickoff for R7 is May 7th PROPOSALS FOR R7 GUILIN FOR WHAT WE PLAN TO BE DOING IN R7:
SUMMARY OF THE STEPS FOR RTCDB "HOW IT OPERATES" (Reference):
A&AI FLOWS: STEP 1...6: Initial A&AI setup of DB (the setup of the DB with the initial set of all xNFs a "getall") STEP 1...6: A&AI Update (e.g. a new xNF is added or deleted) | ||
Renaming the Project | RENAMING THE PROJECT ("Service" vs "Database") Database #1 HISTORICAL PRECEDENCE - The original idea was a configuration database available at Runtime. Use cases to store. Historical been with the project since the beginning. Name Inertia. Operators will use. Historical precedence within AT&T. #2 Contents that it holds - Contents is configuration parameters from the network. Name reflects the initial content of database. Service Since working on project proposal, it has grown, the same argument works against use. #1 QUALIFIERS - A wide variety of qualifiers could be put there and it still won't cover. Would move to something more abstract. Abose and beyond a standard IT database. For example service information, policy information, CLAMP information, exo-inventory (information outside of A&AI), topology information, application information - it is conceivable that many other types of information could before. Config if someone wants to add additional information a place to hold information. e.g. in Bell Canada's case they store more than just configuration, the Operational Data & Current state of network. Collectors that gather metrics in VES consumed put in stateDB. Tied to inventory objects in A&AI self-link from A&AI want to know about interface PNF trying to keep two together, the configuration & the metrics representative what is currently happening in the network. state of I/F being up-down that's more of a state vs a configuration. OpenDaylight Operational data store. Scalability. Collectors & StateDB is yang-driven if collector follows yang-model data store can hold-values. #2 Confederation of Databases - Core/Edge/Far Edge - Historical DB current DB #3 MEANS VS ENDS - Database is a "means" technology not an "end" goal An engine, hubcap is a part of a automobile that provides a service: vehicular motion. A database is a specific technology and implementation. Requirements around for current data & historical (temporal) careful not to talk about the technology. Potentially more than one database. Data Persistency Service → "functional" Data Layer Service Run-Time Configuration DataBase → "technology" State (of Network) Database → what is state of network (storing more than just config) Configuration Operations Database (C.Op.DB) / Swami Golden Configuration Database / Fred RunTime Policy Topology State Configuration Service Exo-Inventory Database | ||
R7 GuiLin Development | CMNotify specification
| ||
Comment made on |
Alessandro Gerardo D'Alessandro |
...