• Main idea: Explore, develop, and share best practices for modeling Vehicle Signal Specification (VSS) data in a NoSQL document database, focusing on performance, scalability, and practical tools that simplify data management for cloud backends.
• Authors: Humza Akhtar@Raphael Schor Rami Pinto 

Introduction

Connected vehicles generate vast amounts of data that is defined by the VSS standard. At MongoDB, we’ve observed an increasing number of customers leveraging the document model to store and query this data in their cloud environments. The flexibility of MongoDB’s schema-less design, coupled with its scalability, makes it a strong candidate for these workloads. Additionally, the VSS tree-like structure maps directly to the document model, providing a natural and intuitive way to organize and manage the data. However, effective use requires thoughtful design choices tailored to application needs and performance considerations.

This proposal outlines an opportunity to collaborate within COVESA to develop tools and best practices that simplify the adoption of document databases for VSS data. This initiative aims not only to provide documentation but also to deliver practical resources—such as migration tools, transformation utilities, and performance experiments—that can help developers and organizations fully harness a document model database capabilities.

Challenges

The design and performance of a VSS data model hinge on several technical and operational factors:

• Network Constraints: Vehicles often rely on unreliable and bandwidth-limited wireless connections.
• Data Transmission Peaks: High data loads tend to coincide with peak mobile network congestion.
• Data Update Frequency: Not all signals are updated at the same rate; some change frequently, while others remain static.
• Workload Characteristics: The system is predominantly write-heavy, with infrequent reads.
• Disconnections: Vehicles may be offline for extended periods, necessitating offline-first considerations.

In addition, the underlying technical stack may introduce performance implications that should inform data model design. For example, considering some technical specifications of the storage engine of MongoDB, WiredTiger, can promote efficient models that minimize document rewrites and leverage the storage engine’s page structure to optimize reads and writes.

Proposal

The initiative will focus on the following areas:

• Data Modeling Best Practices

Develop guidelines for structuring VSS data using MongoDB’s document model. This includes strategies for splitting data across multiple documents to balance performance and maintainability based on common access patterns. For example:

• • Sensor Data (frequently updated) can be treated as metrics and grouped together.
  • Attributes (static data) can be treated as inventory and stored separately.
  • Structural elements that act as placeholders should not be modeled as standalone documents unless necessary.

Specific common access patterns may include:

• • Inventory Queries: Retrieve static attributes such as fuel tank capacity.
  • Operational Data Queries: Analyze metrics such as average fuel consumption over time.
  • Historical State Queries: Track changes in system states over a defined period.

• Tool Development

Create tools that automate common tasks, such as:

• • Data migration from other formats to MongoDB.
  • Transformation and enrichment of VSS data to support various applications.
  • Query optimization utilities based on real-world access patterns.
  • Performance Experimentation: Run experiments to validate modeling strategies and measure their impact on query performance, storage efficiency, and scalability.

• Community Collaboration

Engage with other COVESA members to gather input on access patterns and use cases. This will help tailor solutions that meet the broader ecosystem’s needs.

Goal

The primary goal is to provide the COVESA community with practical, actionable resources that simplify and enhance the use of MongoDB for VSS data. These resources will empower developers to design efficient, scalable, and maintainable data models tailored to real-world access patterns and workloads.

Deliverables

The specific deliverables are open for discussion with the COVESA community to identify projects where this proposal could best fit and provide maximum value. We aim to adopt an iterative approach, delivering tangible results early on.

Our proposed initial deliverables are:

1. A tool that takes VSS JSON as input and outputs a MongoDB schema.
2. A paper detailing the development process of the tool and its potential applications.

Timeline

The timeline for this project includes the following milestones:

• Q4 2024: Preliminary discussions, publish proposal, and define specific deliverables and scope.
• Q1 2025: Complete phase 1 of tool development, present progress at the April AMM, and gather community feedback.
• Q2 2025: Iterate on tooling development based on feedback and publish a paper detailing progress.

Further timelines will be defined as needed based on the final scope of the project and ongoing discussions.

Related resources

• [WEBINAR] Data Modeling in Motion: A live coding session for the mobility sector.
• [BLOG] Accelerate Connected Vehicle Development with Connected Mobility Solution on AWS and MongoDB Atlas.