Versions Compared

Old Version 132

changes.mady.by.user Stephen Lawrence

Saved on

compared with

New Version Current

changes.mady.by.user Stephen Lawrence

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: 18th sept minutes

...

Historical note: the minutes for the previous project, the CVII Tech Stack, can be found here: CVII Technology Stack meeting notes

18th September 2024

Agenda:

  • Fall  AMM
  • Continuation of the Knowledge Layer Connector discussion if BMW are present and there are topics to discuss
  • AoB
    • Ulf: VISSR feeder improvements

Minutes:

  • Fall AMM
    • Steve outlines the Playground session presentation
    • Paul: Are you happy to present the 10min section about experimentation in the Tues workshop?
    • Steve: Yes
    • Richard: Halim will be present at the AMM.
  • AoB
    • VISSR
      • Ulf outlined the work the VISSR project is doing to improve efficiency for subscriptions. Rather than polling, the feeder informs the server when there is a change. See PR for details https://github.com/COVESA/vissr/pull/43
    • AOSP
      • Richard: the AOSP group is looking for ppl interested in working on making VSS signals available to AOSP apps.
      • Steve: Is that mapping to Android VHAL or a parallel server for greater range of properties/signals?
      • Richard: Both of interest.
      • Discussion of the pluses and minuses and some of the prior work.
      • Ulf: VISS could be used for that case.

11th September 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Fall  AMM
  • Knowledge Layer Connector update: BMW want to give an update on their work on the Knowledge Layer components
  • AoB


Minutes:

  • Playground Kanban
    • Steve has fixed issue #57 (online doc ToC not rendered correctly).
    • Updates from BMW - see below
  • Fall AMM
    • Steve and Haonan will discuss Knowledge Layer slides offline
  • Knowledge Layer components
    • Andre reports that BMW have made good internal progress on engineering the knowledge layer components and wants to discuss how to progress publically in terms of PRs etc.
    • Steve suggests following:
      • 1st step would be concentrating on getting the Knowledge layer components public as the info layer ones currently are so we can all see what is being talked about. This could be in the existing PR or a new one if you prefer.
      • Commits can be added as you go to get early feedback. No need for "big bang" single push at the end of development.
      • With AMM coming that gives you some weeks to prepare.
      • Suggested 2nd step would be to consider the modelling of the components, e.g. extend C4 model, to support discussion and review
    • Steve's non-exhaustive list of related topics to be discussed during and after this process as input:
      • Information Layer abstraction
        • data get/set limitations vs data tablet and higher level queries.
          • Need: Agree how to document current limitations and not limit research in higher level and faster queries.
        • database path consolidation for the various containers, e.g. VISS access vs information layer set need to coordinate (now) where to get the data in the DB
          • Need: out of the box components work together
      • Use cases for online documentation
        • Currently intended for more general abstracted documentation and guides that need a lot of cross-referencing, e.g. integration guides
          • Need: Add appropriate doc for the newly introduced features
      • Architecture communication / arch visualisation / C4 model
        • Need: Need arch view of the Playground and how the new features are integrated at different levels of abstraction to be able to discuss within the group and outside.
      • Docker images and packaging of CDSP Core building blocks
        • Pre-built images?
        • Need: Agree on how to handle the dockerisation of Core building blocks in a consistent manner.
    • Discussion:
      • Agreement that current naming of Knowledge and Info layer 'boxes' in current C4 Model are not correct (Steve just added them as placeholders).
        • Collectively need to come up with what is 'better'.
        • Haonan says that Knowledgebase would be traditional way from knowledge perspective but it is also perhaps a little outdated.
        • Ulf suggests "Knowledge Server" as starting point for KL part and both Haonan/Steve agree that is better. Steve changes diagram.
      • Andre happy to add appropriate online documentation
      • Docker
        • Andre wonders about flexibility in selecting docker containers to include when using docker compose.
        • Sebastian says as end user he would simply consume Dockerfiles and deploy with kubernates but sees desire for 'easy' starting point
        • Steve agrees balance needed between out of box experience and both flexibility and maintainability. Docker compose does not preclude someone from really doing some bespoke for themselves.

4th September 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Fall  AMM
  • Playground C4 Model
    • Prior to working on the fall AMM presentation I have been experimenting with various architecture visualisations from the viewpoint of needing to communicate with a wide diversity of job types. The simplicity of the C4 Model looks very promising in the way it combines ease of understanding, useful abstractions and sufficient guide rails.
    • I have drawn a C4 Model of the Playground at the first two levels of abstraction, which you can find on the Data-Arch google drive here and I would like to discuss and get feedback in the meeting.
      • Open the file in draw.io
      • There are 3 pages:
        • L1 Context diagram - Playground System context
        • L2 Container diagram - Major parts of the Playground
        • Speculative draft variation of L2 that adds the Knowledge Layer Connector - I did that just to explore how well the abstraction layering works. You can see for example that containers can be added whilst seperation of concerns maintains existing structures. Also that abstraction allows discussion of the Knowledge Layer Connector interfaces without getting lost in details of SHACL etc.
  • AoB

Minutes:

  • Playground Kanban
    • Steve has merged the online doc Example page changes. Now working on a fix for the ToC indention issue.
    • Did C4 Model diagraming as prep for Fall AMM presentaton. See next.
  • Playground C4 Model
    • Steve presented the Level 1 and Level 2 C4 Model summarised in the agenda above. He feels it shows a lot of promise for communicating the Playground.
    • Main feedback came from Ulf who liked the multi-level abstraction of the approach.


28th August 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
    • Done some work on improving the listing of examples in the online documentation


Minutes:

  • Playground Kanban "stand-up meeting"
    • uProtocol Simulator: Richard will report back if a future strategy emerges from internal GM discussion. Input from the uProtocol community also possible of course. Agreement to place on hold until next steps from developers of the technology becomes clearer. Playground team remains interested if something emerges.
    • Online doc Examples page:
      • Steve shows WIP of updated Examples page that adds the down-sample example and reworks the page to make example discovery a little easier through the use of tables in the categories.
      • Discussion of the WIP. Haonan: a more descriptive heading for the table than "Description" that links the text to the category might be helpful.
    • Knowledge Layer Connector:
      • Haonan requests meeting with Steve to discuss the Connector and AMM
  • Fall AMM
    • Discussion of the Playground presentation and incorporating the KL Connector slides

21st August 2024

No meeting due to holidays.

14th August 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
    • Data Processing PR now merged. Considering what's next
    • I will be on holiday next week. So we either need to cancel next week or someone hosts.

Minutes:

  • Playground Kanban "stand-up meeting"
    • Steve's Data Processing pull request is now merged. He's considering what's next.
    • Halim (GM) reported the team is making progress on dockerising the uprotocol simulator
    • Further commits added to the Knowledge Layer connector PR. Haonan has been working on the RDF Converter.
      • Steve mentions that the IoTDB project is adding Machine Learning support to SQL Queries.
      • Haonan discussed how this could be complimentary to the Rules based Semantic Reasoning. ML could determine some higher value information that is then used by Rules Reasoning to derive the higher level knowledge.
  • Fall AMM
    • Current Introduction session outline:

      • Title:  Introduction to the COVESA Central Data Service Playground

        Short Description: Discover what the COVESA Central Data Service Playground (CDSP) is, what has been done and where it is going.

        Description:   In this session we will introduce the COVESA Central Data Service Playground (CDSP) a project of the COVESA Data Architecture and Infrastructure group.

        • CDSP What / Why / How
        • What has been done since the Spring AMM
        • What’s next / roadmap
  • AoB
    • Steve is off next week. Decision taken to cancel meeting.


7th August 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items

Minutes:

31st July 2024

Agenda:


Minutes:

  • Playground Kanban
    • No change on uProtocol Simulator
    • Knowledge Layer Connector
      • Christian updates on Information Layer component:
        • Further updates in the PR for the IoTDB support in the DB Router.
        • The Router uses Node.js and IoTDB has no Node wrapper for its Thrift rpc API so they have had to work with low level Thrift aspects for get/set.
        • Steve says that IoTDB maintainers are very active on Github Discussions or Slack. Might be worth asking about Node there. Perhaps someone has something.
    • IoTDB timeseries data processing example
      • Steve has pushed substantial updates for the example readme. Walked through readme to discuss structure as possible template for other examples
  • Commercial Vehicle EG outreach
    • Ted explains that the EG is on summer break for two weeks
    • Discussion about data sets, meta-data and collaboration

24th July 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Fall  AMM: https://wiki.covesa.global/x/WgBSBg.
  • First feedback from outreach to Commercial Vehicle EG about their guidelines and meta-data additions.
  • AoB


Minutes:

  • Playground Kanban
    • Waiting for update from GM on uProtocol Simulator.
    • Knowledge Layer Connector
      • Christian has updated the PR to add connection support for IoTDB to the DB Router. No data access yet.
      • Haonan summarised the tasks for the components interfacing Reasoner to Information Layer.
        • Use case for write-back to information layer is when the decision results in values that can be mapped, e.g. driving behaviour results in driver management system alert.
      • Discussion concludes that it is fine for multiple data models to be combined. It's illustrative of a concept, not a prescription as how to implement. Others are welcome to extend with their approach as well.
    • IoTDB Data Quality Processing Example
      • Steve has the data set and the operations worked out. Currently creating the documentation to tie it together.
  • Fall AMM
    • BMW not expected to send several engineers. To be discussed in coming weeks but Steve may present high level view of Knowledge Layer
    • Given attendance of Playground contributors from EU it appears rather than Data Arch specific workshop it may be better to join a more open discussion about scaling and illustrating VSS.
  • Commercial Vehicle EG outreach
    • Steve summarised discussion on their slack channel about their Guidelines and Meta-data:
      • Guidelines have no versioning yet. Location and format may also change.
      • meta-data in the github vspec has no guidelines for how they are be used. Unclear what is required in-vehicle with the data.
    • Further discussion required with them.

17th July 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Fall  AMM: https://wiki.covesa.global/x/WgBSBg. We need to firm up plans as the end of month is approaching. This is in part also a discussion of the group roadmap towards the AMM.
  • AoB
    • Christian: Simulators


Minutes:

  • Playground Kanban 
    • Christian expects to make greater progress on Knowledge connector next week
    • Steve is working on an example for the IoTDB data processing PR. He's got ok from RemotiveLabs to use some of their data.
    • Steve has been contacted by Halim from GM. He is checking the uProtocol Simulator status
    • Knowledge Layer Connector
      • Christian asks about connecting to IoTDB in the Playground
        • During development phase can simply connect to the iotdb-service container running the IoTDB server either from host or another containers as you wish.
        • For deployment we will need to discuss how to integrate the connector, but for example it could be a knowledge-layer image that becomes a service in the playground compose.
  • Fall AMM
    • BMW internal meeting was pushed back. Christian should have news next week.
    • Steve: Chicken and egg on workshops. Holding workshop on playground internals will depend on participation else it might make more sense to join the verticals (commerical, safety, insurance) and discuss use-cases and playground use.
  • Simulator
    • Christian: we have been thinking internally about connecting simulators to the Playground, including the information layer and I wanted to discuss this publically.
    • Discussion:
      • Various simulators will be needed.
      • It's agreed its helpful if only minor glue code is needed to make connections. For that to be possible appropriate documentation is needed for interface integration.
      • Some 'playback' will be as simple as ETL style operations of loading data sets into the DB
      • Standalone collection of data sets reusable across examples is a need.

10th July 2024

Agenda:


Minutes:

  • Kanban meeting:
    • Waiting on GM to return from break for uProtocol simulator update
    • IoTDB Data Processing functions: Steve has updated the PR to add online documentation for the changes.
      • Next is Grafana, plus an example
    • Knowledge Connector: Christian explained updates to the PR, e.g. DB router changes.
      • Discussion in the DB Schema section about challenges of maintaining a single source of truth when relating data models as is typically needed with knowledge layer.
      • For the connector itself it seems ok if it uses a particular approach to communicate the wider idea of using Rules based reasoning. That is internal to the blackbox and other approaches could be used in implementation.
  • DB Schema:
  • Fall AMM
  • Summer holidays:
    • Ulf: off next week and 1 after. Probably some further break later in summer.
    • Sven: 3 week break coming up.
    • Christian: Long break from 8th Aug.
    • Steve: will take 1-2 week break, probably in Aug, to match when others are off.


3rd July 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Docker: Brief summary of my PR above.
  • DB Schema generation: I had some discussion with Gunnar in CVI about this. Basically keeping the meta-data for it with VSS as a layer, rather than dislocated in for example a template language config file. I think ultimately I need to write a fuller proposal that can be discussed, but it perhaps warrants a short discussion if people have input to help form the ideas.
  • Fall AMM: deadline for talks is end of the month. We need to discuss what we want to do, including the possible extra technical day.
  • AoB

Minutes:

  • Kanban stand-up meeting:
    • Knowledge Layer Connector, Christian: Started on skeleten PoC DB router in information layer
    • IoTDB Data Processing, Steve: A WIP PR has been added https://github.com/COVESA/cdsp/pull/52 which adds the optional IoTDB Data Quality Library to the Playground IoTDB docker image and configures IoTDB for access from Grafana.
    • uProtocol Simulator: waiting for update from GM
  • DB Schema Generation:
    • Discussion of providing a mechanism for DB Schema generation for VSS nodes. This would accelerate users ability to more quickly get started with the Playground as it could automate the generation of all nodes.
    • Steve and Christian think this could be useful. Christian in addition raises a similar use-case for the knowledge layer, e.g. SHACL generation. There was some conversion to SHACL in the past.
    • Idea brainstorming:
      • Rough requirements:
        • Input: VSS vspec (a reduced vspec could be input if only a few nodes are required)
        • Generation:
          • Template language + config (config is the description needed to match VSS to the DB)
        • Output: schema for DB X
          • Example: Series of lines to create schema for each VSS node e.g CREATE timeseries.Vehicle.Speed (datatype=INT, compression=X, encoding=Y, meta-data=Z)
      • Steve: In discussion with Gunnar in CVI it was considered that this mapping meta-data could be held in a VSS layer/overlay (closed to the source) rather than in an isolated stand-along config file. If the existing VSS overlay mechanism did not support it, then evolving VSS overlays to be something closer to IFEX layer mechanism could.
    • VSS layer:
      • vehicle.speed DB meta-data=> data-type=INT, compression=squashy, encoding=super-encoding, meta-data="some attribute"
    • Generation / vss-tools:
      • Christian asks about doing this in vss-tools.
      • vss-tools has good set of meta-model functions now, which means generator largely just needs to contemplate output side so could have:
        • vspec2iotdb
        • vspec2realm
        • vspec2shaql
  • Fall AMM
    • COVESA wants technical session proposals by end of the Month.
    • So we need to contemplate:
      • Weds/Thurs Main Conf: Technical Sessions
        • e.g. Outreach etc.
      • Tues Smaller Conf: Workshops
        • e.g. Playground internals
    • Steve will put up the usual page for ideas
    • Paul: COVESA page here All Member Meeting Planning

26th June 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Docker: I have been working on integrating the IoTDB Data Quality library and Grafana. That raises some questions about configuration of Docker containers and the source for the parts we integrate. That may be worth a discussion to get some feedback. Particularly from the BMW team who may be doing similar work at the moment.
  • @Swen Schisler (Endava) mentioned last week wanting to discuss some of his proposal to the CVI group
  • AoB


Minutes:

  • One possible set of container images for the Playground compose:
    • Image1: iotdb
    • Image2: VISSR
    • Image3: playground-tools/extras
      • Knowledge Connector
        • Info Layer
          • Realm backend widget
        • Knowledge

19th June 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • GM uProtocol Simulator update: two weeks ago GM said they would update us on progress of dockerising the simulator. This would also be an opportunity to advance the integration discussion.
  • Continuation of the two topics @Ulf proposed last time and briefly introduced
    • Playground documentation structure
    • Proposal for integration of VISS into the Knowledge Layer connector
  • AoB


Minutes:

  • Playground Kanban stand-up
    • 0.1.0 blog will be published any day now.
    • Steve reports that he has introduced COVESA and the Playground project to the Apache IoTDB community. Received positive interest in the Playground and willingness to support IoTDB.
  • GM were scheduled to update us on the uProtocol Simulator but no GM representatives attended.
  • Continuation of the two topics @Ulf proposed last time and briefly introduced
    • Playground documentation structure
      • Ulf described his proposed updates to layout of the landing page and organising the examples.
      • Screenshot: Ulf-CDSP-doc-layout.png
      • Steve described original was modelled after the Docker landing page which does a nice job of enabling both newcomers and experts: https://docs.docker.com/
      • Ulf's point about seperating the user/ref manuals of the core components would be useful
    • Proposal for integration of VISS into the Knowledge Layer connector
      • Agreement that the pluggable architecture in the WS component for DBs would allow Ulf to attach some of his idea, e.g. he feels he could use OVDS to attach VISS in the cloud.
      • Discussion of other aspects
  • AoB
    • Swen mentioned he could discuss some of his ideas for his project next week

12th June 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • @Christian Mühlbauer @Haonan Qiu will present their idea of a Knowledge Layer connector example
  • AoB
    • @Ulf has proposed two agenda items today:
      • Playground documentation structure
      • Proposal for integration of VISS into the Knowledge Layer connector
    • If there is time at the end of the meeting we can get an overview with a view to progressing discussion next week and or in slack

5th June 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • uServices mocker integration discussion with GM team
  • Time permitting or uServices discussion does not proceed: Knowledge Layer connector example
  • AoB

Minutes:

  • Playground Kanban stand-up
    • Steve has made the 0.1.0 release. Currently working with COVESA marketing to publicise
  • uServices Mocker discussion
    • Halim set the big picture. Mocker in question is the Eclipse uProtocol Simulator. Mocker seems a good fit for the concept of the Playground.
      • From project README.md: The up-simulator is a user-friendly web tool designed to make developing uProtocol services and Apps on PCs easy. Its intuitive interface empowers users with features such as Publish-Subscribe, RPC invocation, and the simulation of COVESA Vehicle Services. With a focus on a user-friendly graphical interface (GUI), users can easily navigate through the development and testing phases of uProtocol Apps, ensuring a smooth experience from start to finish.
    • GM engineers demonstrated the simulator for uProtocol and vsomeip traffic.
    • Halim: whilst the simulator supports uProtocol and uServices it can be used to simulate pure SOME/IP traffic.
    • Steve: as the simulator already exists it seems to be a question of how the 'tool' would be integrated and connected to other components to add value. Is it containerised?
    • Simulator is implemented in Python, but not yet containerised.
    • Halim suggests that GM containerises the simulator and that is then used to guage wider interest. This seems like a sensible way forward.
    • Halim: there will be a company stop in around a month for a summer break, so I would like to push the GM team to make some progress in the next two weeks. Let's plan for a progress update from the GM team in the Data Arch call for 19th June.
  • Knowledge Layer Connector example
    • Ran out of time. Pushed back to next week.

29th May 2024

Agenda:

  • Playground Kanban "stand-up meeting": In-progress task status, major new items
  • Knowledge Layer connector
  • GM uServices mocker meeting next week
  • AoB


Minutes:

  • Playground Kanban
    • After discussion with Arnaldo decision taken to move https://github.com/COVESA/cdsp/issues/15 back to the backlog as its currently more narrowly defined as being about Realm as a backend to VISSR. Work on Realm support for other use is currently a part of the Knowledge Connector.
    • Steve has started on the 0.1.0 release. After that he may continue with the work on Grafana and IoTDB data quality functions.
    • BMW have started public work on the Knowledge Connector in their fork.
  • Steve has confirmed with Halim that we will discuss integration of the GM uServices mocker into the Playground in next weeks call.
  • AoB
    • Paul asks for input on marketing projects for outreach
      • Steve: I have been talking to Mary about utilising the COVESA Blog and social accounts. General idea is to semi-regular post updates as new features added or releases made.

21st May 2024

Agenda:

  • Playground Kanban: In-progress task status, major new items
  • Knowledge Layer connector: code directory layout
  • CVI call to discuss GM uServices mocker
  • RL meeting
  • AoB

Minutes:

  • Began Kanban stand-up meeting.
  • Pivoted into longer discussion about Knowledge Layer Connector
    • Ulf and Christian discussed VISS relationship
    • Discussion of code layout and the components. Christian showed update list of components and their functionality.
    • Discussion of upcoming code commits from BMW. They will create a PR. Possibly starting with READMEs for the various parts
  • Ulf discussed earlier work in CCS OVDS database in cloud. Thinks it can be reused.
  • Steve will confirm with Halim about when the discussion with GM team regarding integrating the uServices Mocker will take place and report on Slack
  • Steve outlined positive meeting he had with RemotiveLabs introducing the Playground to them. Follow up required on areas of collaboration.

15th May 2024

Agenda:

  • Playground Kanban: In-progress task status, major new items
  • 0.1.0 release: I would like to make an initial release soon, but would like opinions if anyone has input on what/how
  • Knowledge Layer connector
  • Backlog generation
  • AoB: related projects etc.

8th May 2024

Agenda:

  • Playground Kanban
    • In-progress task status
    • 0.1.0 release
    • Completing project setup
  • Backlog generation
  • AoB


Minutes:

  • Playground status
    • Steve has a WIP PR to add the RemotiveLabs feeder/bridge he created for the Spring AMM. Documentation is next major step
    • Christian presented latest thinking on the Knowledge Layer connector. Extensive discussion of that.
    • Steve: we should look to make 0.1.0 release soon and complete project setup
  • Limited timing remaining but some discussion of backlog generation and the need for use cases for data processing/analytics
  • AoB:
    • Ulf:
      • VISS/VISSR weekly project meetings starting Monday at 4-5pm CET
      • Commercial Vehicle Information Specifications project under Commerical Vehicle BoF being scoped. Starting with Tractor/Trailer trees.

1st May 2024

Minutes:

  • May national holidays in EU meant missing active members so informal discussion held.
  • General discussion of data sets, data processing and need for functional requirement input from industry.

24th April 2024

Agenda:

  • AMM feedback
  • Team backlog

...

  • Backlog
    • Playground backlog slide from AMM:
      • Next (goal: enhance and use the playground):
        • Review 0.1.0 release and fill obvious doc gaps
        • Socialisation: COVESA groups, Webinars etc.
        • Concrete backlog generation (high lvl points below)
          • Main focus of Thurs Data Architecture team workshop
          • Feeders: VISSR feeder component, RemotiveLabs, etc.
          • Examples: agree data arch team early focus.
          • Connections to COVESA eco-system, e.g. CVI southbound, AOSP northbound
          • Embedded App DB (Realm)
          • DIKW Knowledge layer connector
          • Auto h/w deployment (mid term)
          • Complete project setup: release process etc
  • Discussion in meeting:
    • Need data sets (at rest):
      • VSS data sets (real or simulated)
      • Other data models and rule sets.
    • Next week or two Steve will do some documentation polishing and publish the RemotiveLabs bridge sample code as an example (source is available in his github so ask if you want to know sooner)
    • Discussion of knowledge layer
      • Christian/Haonan are continuing to work on the connector
      • Discussion of rules example Haonan posted to the slack.
      • She explained she's currently bypassing some steps in the processing chain, but is just directly looking at the applicability of processing the timeseries data in RDFox to confirm approaches.

...