Connected Vehicle CyberSecurity Research Group at MIT CSAIL

Goals for this presentation

• Form a Connected Vehicle CyberSecurity Research Group at MIT
• Increase industry awareness of W3C Auto activity

Topics

• Brief overview of W3C
• Summary of W3C Automotive and Security Activities
• Research Group Proposal
• Next steps

Overview of W3C

• Standards body for the Web
• International consortium founded in 1994
• Director Tim Berners-Lee
• Headquartered at MIT
• >480 Members
• Approach standards from individual technology and Industry focuses

Why Web standards for Auto?

• Strong need for a robust application ecosystem
• Reduce integration nightmares, avoid vendor lock-in (RFPs)
• Reduce fragmentation by competing proprietary approaches
• Attract Content Providers - 3rd party apps
• Enable on-board data sampling/event triggers for silos and v2x
• Web has proven itself as transformative for multitude of industries
• Connected vehicles relying on Web services
• W3C history of providing higher level interfaces to developers

Convincing Content Providers

• Multitude of competing platforms
• Compare to Smartphone Market
• App developers unwilling to supporting many platforms
• Open Source ≠ Open Standards
• Need for commonality, holding up progress
• more than just user apps - building block for ADAS & autonomous, can be part of solving data sampling needs for silos, regulators, V2x, insurance...

Primary stakeholders

• Auto Manufacturers
• Tier 1s
• Payment providers
• Fuel and service companies / Power utilities
• Telcos
• Insurance Companies
• Fleet Management
• Solution and Content Providers
• Federal, State/Province and Municipal Governments

W3C Automotive Activity

Creating a rich application ecosystem with vehicle service specifications

• Vehicle Signals (v1, v2 will converge with VW Group)
• Media Tuner
• Location Based Services
• Automotive Web Payments
• RESTful Service Interface
• Qt, HTML5 or headless in practically any programming language

Automotive Security & Privacy

• Authentication and granular access control considerations built into specification
• W3C Automotive Privacy and Security Task Force
• Genivi Security Expert Group
• Conversations, workshops on forming Research Group at MIT
• Liaisons with various W3C Privacy and Security groups
• Coordination with various organizations

W3C Security

Security and Privacy on the OpenWeb Platform

• WebAuthN Working Group
• Web Application Security Working Group
• Web Crypto API
• Privacy Interest Group
• Web Security Interest Group

Relevant Work at W3C

• Vehicle Information Service Specification (VISS)
• Vehicle Information API Specification (VIAS)
• RESTful Service Interface (v2 based on VW submission)
• Location Based Services
• Automotive Web Payments
• Efficient XML Interchange (EXI)
• Second Screen Working Group
• Near Field Communication Community Group
• Web Accessibility Initiative for possible Distractability concerns

Relevant Work at W3C (cont)

• Web Sockets
• Speech API Community Group
• WebAuthN Working Group
• Web Application Security Working Group
• W3C Crypto API
• Privacy Interest Group
• Blockchain Digital Assets Community Group
• Web of Things
• incomplete list, new pertinent activities forming

Get Involved in W3C Auto standards!

• Influence on use cases, requirements and prioritization
• Earlier adoption and insight leads to faster testing and product implementations
• Collaborate with new industry players for new opportunities
• Access to new expertise; enhance skills of current employees

Problem Definition

• Automotive application framework fragmentation - challenge & opportunity
• Content providers reluctance
• Internet is connected vehicles' biggest attack vector

Internet / connected vehicle attack vector

Allowing third party applications to interact with the underlying vehicle and services from the Internet provides a rich environment but also rife with risk.

Internet / connected vehicle attack vector

• Autonomous and ADAS decisions are data driven
• External data can be manipulated, have malicious content interjected and be delivery mechanism for attempts to execute arbitrary code

Potential Solution

Form a research group at MIT Computer Science and Artificial Intelligence Labs

• Separate activity at MIT, compliment to W3C Auto work
• It would contribute towards W3C Privacy & Security Guidelines
• Network connectivity regulation
• Data Integrity System
• Potential areas of exploration open to input from sponsors

W3C Guidelines for 3rd party IVI applications

• Clear, enforceable rules for network and application interactions
• Application resource package manifest
  Specific examples in Genivi Security Expert Group presentation
• External content permission policies
• Privacy policies (eg GPPR) for sharing vehicle information with external services
• While these will be written with W3C Auto standards in mind, goal would be to make these platform agnostic

Data Integrity System

• monitor all connected vehicle application information exchanges
• adaptive, relaxing or restricting, from rulesets and machine learning
• external retrieved content type restriction
• malicious content checks (eg tainted images)
• ensure volume and frequency metrics adhere to expectations
• perform data sampling, look for egregious anomalies
• control what data is being sent off vehicle
• blocking trackers
• provide caching for performance and sporadic connectivity concerns

Methodology

• Evaluate vehicle-to-everything (V2x) data exchanges
• Define a full attack tree, find weaknesses and devise mitigations
• Leverage and expand findings from Genivi Security Expert Group and W3C Automotive Privacy and Security Task Force
• Establish information exchanges and collaborations with other researchers

Methodology

• Test vehicle / mobile lab / surveillance van
• Conduct field studies - sniff "traffic" see how much is leaking
• Setup Genivi Development Platform, AutoGrade Linux reference implementation and other BSD or Linux based Head Units
• Explore V2x vulnerabilities and means to modify or inject malicious data
• Prototype Data Integrity and Guideline Enforcement Systems

Potential Deliverables

• Threat Model for IVI, Applications and external website interactions
• W3C Privacy and Security Guidelines
• Study findings
• Connected Vehicle Application Information Exchange Integrity System
• Network focused IVI Penetration tests
• External website security assessment evaluation tools, automated/recurring and manual
• App testing, rule generation and review service tooling and evaluation workflow for integration into app marketplaces

Potential Liaisons

• OEM and independent security research groups
• AutoISAC
• Genivi
• Linux Foundation Automotive Grade Linux
• SAE
• Autosec.org
• US DOT Volpe, UK DfT and other government entities
• Member organizations participating in W3C Automotive activity

Requirements

What is needed to form this group

• One or more OEM - can't afford to brick vehicles
• CSAIL Alliance requires 4 supporting sponsors and/or a significant seed sponsor to start
• Input from stakeholders

Sponsorship Benefits

• Influence research scope
• Early access to findings and proposed W3C Guidelines
• Access to PoC evaluation, enforcement and integrity systems
• Submit platforms for evaluation
• Recruitment opportunities for MIT graduate students participating in this research project
• Access to other Cybersecurity@CSAIL (MIT) research projects
• Potential for partnerships in spinoff ventures

Thank You

• Proposal
• Questions
• Follow up: ted+auto@w3.org
• W3C Auto overview slides https://www.w3.org/auto/slides