...
- First/early 'next steps'
- How to organise project during its creation phase?
- Move discussion to 24/7 online where possible for faster pace than weekly calls
- Top level phases?
- Record and comment on work breakdown.
- Project management: where, what, how
- Take a breath - are we aligned on top level goals and what/who/how? Who will be active? What are their interests?
- Implementation
- Check state of VISS Docker support
- Confirm project setup for upstreams, e.g VISS, DBs
- Can be considered part of early wider socialisation of project.
- Planning for additional DBs, e.g. Apache IoTDB, Realm
- Determine early Use Cases to be 'guided' by
- Look at connection to MongoDB/BMW work. What is common?
- Feeder link
Use cases to be explored
tbd
...
Phase Planning
Phase 1 Create Early PoC
High level content:
- Playground
- WAII VISS
- Connect Realm DB to WAII
- Connect Apache IoTDB DB to WAII
- Docker deployment of Playground
- Scope: starts container chain containing above
- Documentation website
- Scope: online instance (prove website framework has what we need at high level), with some representative (possibly placeholder documents)
Concept drawing in DrawIO: https://drive.google.com/file/d/1hxxDVgqianoMoXapPsYfjKgQVdc6Uume/view?usp=sharing
Best Practices
List of best practices that may become a best practices document for the Playground
- Docker
- Official Docker project best practices:
Project Management
In 25th Oct 2023 meeting (see minutes for details) following decisions were made:
- PM task tool
- Decision: Start with github kanban for project management
- Comms
- Decision: Covesa Community Slack. Create a new data-architecture channel.
- Phases
- Discussion favours doing simple early PoC code to help decide on source and project organisation.
- Decision: Needs refinement in the kanban board but first phase is 'create Playground PoC' along lines above.
Oustanding questions:
- How to handle Epics (dev themes) which is not a feature of Github Projects? e.g. use undated Milestones.
- How to handle tasks in a ticket that are not directly development such as operational tasks? Could use enhancement label then have problem of filtering those from feature enhancements.
Use cases to be explored
Christian Muehlbauer Input (Proposal):
User Story:
"As a user of the Playground, for example, I would like to be able to quickly implement the following use case - representative for a Use Case “touching all touchpoints” of a data-centric automotive ecosystem - in a fast proof of concept implementation using Lego pieces from the Playground."
Validation Use Case:
(Based on: https://wiki.covesa.global/pages/viewpage.action?pageId=71074417)
Image Added
Ideally, the Playground Repo provides a selection of components for each „X“ for the corresponding target platform (SDK, or docker, etc):
Image Added
Steve input (brainstorm list):
- General
- Some aspect of the BMW / MongoDB project in the open
- Look at MongoDB digital twin example
- Interfaces: Common Service definitions, IFEX, Autosar VehicleAPI
- Data Expert Group touchpoints
- Internals
- Medium and high speed data, Latency etc. as previously discussed in the group
- Southbound
- vSOME/IP example
- RemotiveLabs feeder
- Carla feeder
- Northbound
VISS Protocol Servers
Notes on known OSS VISS Servers:
...
- Apache IoTDB
- Column orientated time-series DB designed for embedded IoT (high ingestation rate, low latency query etc.)
- User Defined Functions for data processing at the Edge
- Wide support for Apache eco-system, e.g. Flink, Hadoop, Spark in the cloud.
- Wide range of API bindings + MQTT broker
- Some sync functionality between instances
- Possibility to use simplified TsFile API in low resource environment, although currently there is no C, C++ or Rust implementation.
- MongoDB Realm
- Object DB initially developed for mobile.
- Wide range of mobile API bindings + C#. C++ binding in Beta.
- Advanced support for automated sync between instances. Also sync between Realm and Atlas (MongoDB Cloud DB) in the cloud.