Agenda July 25th (Virtual),  14:00 -18:00 CET / 8am - 12pm ET / 5am - 9am PT

Antitrust
Before we begin, we would like to make clear that COVESA is committed to compliance with the antitrust laws in all of its activities, and that it expects all participants to similarly comply with the antitrust laws.  We will not engage in--and members must refrain from--any discussion of, or understandings regarding competitively sensitive topics. If you have any doubts regarding whether a matter is appropriate for discussion, please consult with your antitrust counsel.

Open and Royalty-Free
Further, COVESA aspires to be an open and royalty-free organization. The discussions and contributions made during this session are governed by the COVESA Intellectual Property policy. If you are unfamiliar with that policy, please review it in detail prior to making any contribution that reads upon a patent.

Presentation:

COVESA_VSS_OpenAPI.pdf

Participants:

Image Added

Late joiners:

• Richard Fernandes
• Achim Henkel

Agenda and MoM

• Welcome and introduction (10 minutes)
  • Paul introduced the background
• Round the table - how do you use APIs today?, what are your general expectations
  • Paul: Would like to hear your thoughts
  • Gunnar: My focus is on implementation/results, not aligning all OEMs, willing to discuss actual contributions. Intends to do a presentation on what has been done any why.
  • Adnan: Has prepared an overview of how things can be done in standardized way, but most important to get alignment on IFEX
  • Halim: Clarify/Align on strategy of interface pillar. Like the separation between IFEX (syntax) and catalog (VSC). Would really like to see a catalog
  • Achim: Interesting in how we can make use of VISSv2 - do we need a VISSv2 light?
  • Neil: VSS successful, but standardized way of exchanging data would be good, no need for separate integration effort
  • Tim: We see a possibility for standardization on lower levels, like zone controllors, device abstraction layers. For example, good if you can control seat heater plates.
  • Erik: In some areas like wipers we already touch "lower levels" with "system" signals
  • Stephen: Wants anti-fragmentation. Interesting in capability discussion, service orchestration
• BMW Idea: OpenAPI/AsyncAPI presentation
  • Concerns Zone ECU to Integration layer
  • Focus on Client-Server and event driven
  • Adnan presented API example
  • EriK: So you can use VSS as catalog, and then let vspec2jsonschema have attributes/variants to for example support history read, timestamps to be included
  • Adnan: yes
  • Adnan: AsyncAPI include queuing concepts
  • Mariusz: How is deployment described
  • Adnan: Have tested both integrated and separate files
  • Gunnar: AsyncAPI supports pub/sub, OpenAPI more for REST. In AsyncAPI some deployment details, like channels, but they are described in abstract manner. In IFEX even strict(er) split between interface description and transport methods / deployment
  • Adnan: You can specify

• Gunnar: Update on IFEX and VSC
  • Presented IFEX as "common" denominator when moving from one API to another. - IFEX-quick-what-and-why-v2.0.pdf
  • Erik: Do we want "IFEX" (whatever that is) to be a subset of most other IDLs, or rather a superset, both gives challenges
  • Gunnar: It is doable to do something that converts from "expressive" to "simple". Like if the target cannot handle constraints you drop it, perhaps the tool give warnings if certain details are dropped.
    
  • Gunnar: ARXML and D-Bus prioritized areas
• GM/Halim: GM collaboration and contribution to Vehicle Services
  • Halim presented  uServices - SOA 2.0, Resiliency, ...
    
    • uProtocol, Protobuf
  • GM interested in contributing to a "uServices" catalog.
  • Possibly also a Mock service implementation and BDD feature Files
  • We have a vSS→Protobuf. Rely on protobuf custom options, can annotate topics we want to publish.
  • VSC / IFEX License discussion. Is it possible to switch to Apache?
    • Erik: We need guidance on the process from the board. So the board can make the decision, as the board will be the one that will be in problem
    • Gunnar: Will do some investigations (for IFEX)
      
• RTI: Expanding VSS beyond present footprint
  • Neil: A lot of domains, would it be possible to extend scope outside telematics.
  • Current scope limits the reach
  • There is an integration effort, could COVESA extend the scope. Control-signals, Alarm-signal
  • Create datatypes that match datatypes used in the industry
  • Improves open source story
  • Could possibly be a common language for others like SOAFEE.
  • Sebastian: Is it only a model question - more signals but on lower levels, or do we miss features
  • Halim: GM wants to see service catalogs with those types
  • Sebastian: Should this be done like EV charging, a group that spins off
  • Stephen: Need to be careful to make a distinction between VSS catalog and VSS syntax
• Conclusion - next steps
  • Achim: For the RTI proposal, would like to see a slide/deepdive
  • Halim: Want to work on the catalog part
  • Gunnar: There is Mercedes (long term) interest in catalog
  • Erik: Having some realistic services defined in VSC/IFEX would be great. Could it be an idea to try to find a few "areas" for services and see if multiple organizations are interested in any of those areas.
  • Halim: We could look at standardized VHAL properties
  • Sebastian: If you have some services already defined, it is good to put it "out" to get Feedback
  • Halim: Plan for October
  • Let us discuss hands on activities and next meetings on DEG meeting on Thursday this week
• Topics below not covered
• Target: Agree on what is feasible to focus on in the short term
  • What is possible to achieve until Spring AMM
• Way forward
  • F2F September - next to ETAS connection?
    • If so - time and agenda
  • Session at AMM (Monday?)
• Discussion on work-areas for the interface pillar. Target is to identify the work-areas and how important they are considered by participating organizations. Some possible work-area proposals below
  • Common IDL
    • Is there a value in agreeing on a "common IDL format" (existing or new) that could be used for describing services?
    • If so, why? Some possible reasons:
      • Foundation for a common catalog
      • Foundation for implementation of reference cliensts/servers/SDKs
      • Simplifies translation/mapping between other IDLs,
  • IDL translation capabilities
    • What type of translations are of interest?
    • What do you need? Just a specification (like ARXML to Protobuf), or actually also tooling?
    • What are you willing to work-on?
  • Common service catalog
    • Do you see a value in a common catalog for services (like for example FOTA, diagnostics, ...)?
    • Do you think it actually is feasible? What is needed for success? The catalog part of VSC has so far not really been any success!
    • Do you see any candidate areas where chances of success is bigger than for others?
  • Service Capabilities as an abstraction mechanism (input from Stephen Lawrence )
    • MBition, BMW and GM have all expressed interest in defining interfaces (in a technology agnostic manner) that describe 'basic' capabilities (e.g. move seat to X) on which value added services are built and can be reconfigured.
    • Just as VSS can be considered an abstraction mechanism for data, interfaces (VSS+methods) describing Service Capabilities could be considered an abstraction mechanism for implementation details below. Abstracting away 'small ECU' real time or ASIL implementations for example.
      
    • Do we see value in discussing Capabilities as an abstraction mechanism? Two potential topics come to mind:
      • Grounding discussion in likely deployment scenarios. Where they sit in the stack and what is around them to meet the design goals. I mean that largely as a verticality discussion in the stack as to what it may be interfacing too (e.g. Vehicle API) rather than a transport, protocol or system architecture selection discussion. In this way powerful, but overly complex approaches can be avoided if the problem space is simpler. It also naturally links discussion to other work streams in Covesa such as Vehicle API and data architecture.
      • Grouping Capabilities into Catalogs - this is related to the Service Catalog topic as the interfaces for the capabilities may be bundled into functional groups that form the basis for a catalog that is either unique to the provider (internal + trusted partners) or in its more open form a cross industry catalog. I list the topic separately from a Common Service Catalog as if a decision to not pursue a Common Catalog is made Capabilities remains a useful organising concept.
  • Reference implementations
    • Do you think COVESA shall be active doing reference implementations for Client/Servers/SDKs? Or is that better handled by other downstream projects (like W3C and Eclipse SDV)
      
  • VSS relationship
    • Shall COVESA define one or more APIs (using the "Common IDL") for accessing VSS data?
      • As a real API definition, or just as requirements on an API definition. Potential requirements include:
        • Support for atomic read/write/subscribe of multiple signals
        • Support for both "south" (the one providing values or actuating requests) and "north" (the one reading values and requesting actuation) i.e. support both for reporting actual value and for setting wanted value.
        • Metadata to be provided (like timestamp)
        • Error codes or scenarios - why was an actuation not fulfilled?
    • If so, focusing on thin API (like VISS and KUKSA.val) or fat (one method per signal)
    • Need for possibility specifying services that references VSS-signals, for example service to set 3 VSS signals in a single call
  • Relationship to COVESA-AUTOSAR Vehicle API discussion
• Target: Agree on what is feasible to focus on in the short term
  • What is possible to achieve until Spring AMM
• Way forward
  • F2F September - next to ETAS connection?
    • If so - time and agenda
  • Session at AMM (Monday?)

...

...