In order to drive discussions on security / access control, sensitivity in terms of privacy, and many other aspects, it is important to be able to categorize vehicle data in groups (including multiple and overlapping views / filters / layers).   Analysis of this activity is started on this wiki page

In the GENIVI Technical Summit (November 2019), it came to discussion, how different vehicle datasets could be defined in order to provide OEM guidelines.

"To define datasets that can be provided for specific purposes/use-cases to have consistent coverage from OEMs (bundling of data per use-case)"

It should be kept in mind that different services have different data needs, and might need an expanded dataset in order to function. Still, sorting data points into different bundles makes the data use cases clearer compared to listing individual data points. Apart from these first data bundles, we should also consider different categories based on data sensitivity, privacy and regional laws.

Proposed Definitions

• Data Category = Data that belong together as part of a common technical scope.  
  In the tree structure of VSS we see this natural subdivision using branches.  In VSS a lot of the structure is mirroring the physical structure of a car, but occasionally other things like the organisational structure of automotive development companies or other divisions have also been influencing the category hierarchy.  Categories could include larger groups that share a particular common characteristic, e.g. "Privacy sensitive data".
• ExVe Container = Grouping together previously unrelated data for a new purpose or to support a certain function or use-case (which suggests that ExVe Container definitions are frequently created, possibly short-lived, possibly locally defined, etc.)  The Extended Vehicle ISO standard speaks about "Containers" for this purpose.   We prefix the term with ExVe because "container" is a very frequent and overloaded term.
  • Example:  A pay-as-you drive insurance application may need approximate positional data (approximate for privacy) and the odometer data and some engine usage parameters.  Those would normally 
  • Such containers could also group together data that affect access permissions, or a logical group of information that a user gives consent to share.

Most services need data points from across different technical categories. E.g. a service might need charging.state_of_charge and engine.oil_temperature. It could be said that they need to access two different data categories. It could also be said that they need to access a "Container" that includes these two different data points. To build up consistency inside companies, but also across, it makes sense that common use-cases have pre-defined containers. These are typically called "Data bundles" or "Data buckets", which is like a template of a container that has other meta information attached to it like purpose of use, pricing and rate limits.  

(When we speak about transferring measured values that are bundled together, note the related definition of a Snapshot)

Example Categories and Containers:

Personalised vehicle data – read-only

API access – write