You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Over the past two years, the Google Ad Traffic Quality team along with other teams at Google have experimented with the Trust Tokens API and Android Platform Provided Trust Tokens to defend against invalid traffic (IVT). The team found that the API appears to be better suited for authentication purposes (as illustrated by its use in the Privacy Sandbox k-Anonymity Server) rather than for IVT detection. To use the API for IVT defenses, the Google Ad Traffic Quality team recommends key improvements be made to enhance coverage (as described below).
We found several limitations in being able to use Trust Tokens for IVT detection. We need to maximize availability of trust tokens wherever ads are shown to maintain the integrity of our defenses against IVT. We suggest the following:
Allow token redemption and request signing in cross-origin frames by not requiring the trust-token-redemption Feature Policy enablement.
Enable trust token use on HTTP domains as ads continue to be served there.
Adjust the limits on the number of issuers and redemptions to account for limits we will likely hit after widespread adoption of the API and lead to reduced coverage.
Debuggability Feedback
The API was easy to use with iframe/fetch/xhr, and the debuggability was great with DevTool support, but the API would further benefit from the following:
Richer error states and messages. For example, sending a redemption record irrespective of whether one exists or not should not lead to errors. But, occasionally we see “XHR Error.” Lack of clear error states and messages makes it hard to identify the root cause of the error.
The ability to detect scripts that override fetch (or other APIs) would help identify the gaps in trust tokens coverage.
Summary
Over the past two years, the Google Ad Traffic Quality team along with other teams at Google have experimented with the Trust Tokens API and Android Platform Provided Trust Tokens to defend against invalid traffic (IVT). The team found that the API appears to be better suited for authentication purposes (as illustrated by its use in the Privacy Sandbox k-Anonymity Server) rather than for IVT detection. To use the API for IVT defenses, the Google Ad Traffic Quality team recommends key improvements be made to enhance coverage (as described below).
Scale of Experiment
The team experimented with the Trust Tokens API on 100% and Android Platform Provided Trust Tokens API on 5% of the traffic for publishers with Google’s ad tags from Chrome instances supporting Trust Tokens Origin Trial (OT).
Coverage Feedback
We found several limitations in being able to use Trust Tokens for IVT detection. We need to maximize availability of trust tokens wherever ads are shown to maintain the integrity of our defenses against IVT. We suggest the following:
trust-token-redemption
Feature Policy enablement.Debuggability Feedback
The API was easy to use with iframe/fetch/xhr, and the debuggability was great with DevTool support, but the API would further benefit from the following:
fetch (or other APIs)
would help identify the gaps in trust tokens coverage.Value of the APIs for IVT Detection
Both the Trust Tokens API and Android Platform Provided Trust Tokens API were found to be helpful in identifying IVT, but we anticipate that the Cross-Partition Cookie Age Signal and Device Integrity Attestation APIs may be easier to use and probably more beneficial for IVT use cases like the ones described in this comment.
(Note: For confidentiality reasons, we are unable to share the specific examples of IVT or the attack vectors these APIs would help against.)
The text was updated successfully, but these errors were encountered: