LGTM to experiment till M104, inclusive

On Thu, Mar 31, 2022 at 1:38 PM Paul Jensen <pauljen...@chromium.org> wrote:

>
>
> On Wed, Mar 30, 2022 at 5:53 AM Yoav Weiss <yoavwe...@chromium.org> wrote:
>
>>
>>
>> On Friday, March 25, 2022 at 11:46:45 PM UTC+1 Paul Jensen wrote:
>>
>>> Contact emails
>>>
>>> pauljen...@chromium.org <https://groups.google.com/>, kle...@google.com
>>> <https://groups.google.com/>
>>>
>>> Developers interested in the FLEDGE API can also join the FLEDGE API
>>> announcements
>>> <https://groups.google.com/a/chromium.org/g/fledge-api-announce> group
>>> for updates and announcements.
>>>
>>> Explainer
>>>
>>> https://github.com/WICG/turtledove/blob/master/FLEDGE.md
>>>
>>> Specification
>>>
>>> May be heavily influenced by origin trial feedback, so not yet started.
>>>
>>> Summary
>>>
>>> FLEDGE provides a privacy advancing API to facilitate interest group
>>> based advertising.  FLEDGE shifts the interest data and the final ad
>>> decision browser-side instead of server-side, offering many advantages:
>>> strong privacy guarantees, as well as time limits on group membership,
>>> transparency into how the advertiser interest groups are built and used,
>>> and granular or global controls over this type of ad targeting.
>>>
>>> Blink component
>>>
>>> Blink
>>> <https://bugs.chromium.org/p/chromium/issues/list?q=component:Blink>
>>>
>>> TAG review
>>>
>>> https://github.com/w3ctag/design-reviews/issues/723
>>>
>>> (The first “F” and last “E” in FLEDGE stand for First Experiment, as
>>> such FLEDGE specifies a prototype for ads serving in the TURTLEDOVE family,
>>> so the TAG review request was made for the overall family, TURTLEDOVE,
>>> rather than the first experiment specification, FLEDGE.)
>>>
>>> TAG review status
>>>
>>> Pending
>>>
>>> Risks
>>> Interoperability
>>>
>>> Gecko: No signal
>>>
>>> WebKit: No signal
>>>
>>> Edge: Edge is also exploring interest group based advertising, namely
>>> with the PARAKEET proposal
>>> <https://github.com/WICG/privacy-preserving-ads/blob/main/Parakeet.md>.
>>> PARAKEET shares much of its API with FLEDGE but has a different trust
>>> model.  Deployment experience is necessary to inform the choice between the
>>> trust models.
>>>
>>> Web developers: Significant interest from many web advertising
>>> technology developers.  WICG FLEDGE calls
>>> <https://github.com/WICG/turtledove/issues/88> are heavily attended.
>>> Interest in FLEDGE is further evidenced by the many related discussions
>>> and proposals that FLEDGE design draws from, most notably:
>>>
>>>    -
>>>
>>>    The original TURTLEDOVE
>>>    <https://github.com/WICG/turtledove/blob/main/Original-TURTLEDOVE.md>
>>>    from Chrome.
>>>    -
>>>
>>>    SPARROW <https://github.com/WICG/sparrow> from Criteo.
>>>    -
>>>
>>>    Outcome-based TURTLEDOVE
>>>    <https://github.com/WICG/turtledove/blob/main/OUTCOME_BASED.md> and 
>>> Product-level
>>>    TURTLEDOVE
>>>    <https://github.com/WICG/turtledove/blob/main/PRODUCT_LEVEL.md> from
>>>    RTB House.
>>>    -
>>>
>>>    Dovekey
>>>    <https://github.com/google/ads-privacy/tree/master/proposals/dovekey>
>>>    from Google Ads.
>>>    -
>>>
>>>    PARRROT
>>>    
>>> <https://github.com/prebid/identity-gatekeeper/blob/master/proposals/PARRROT.md>
>>>    from Magnite.
>>>    -
>>>
>>>    TERN <https://github.com/WICG/turtledove/blob/main/TERN.md> from
>>>    NextRoll.
>>>
>>>
>>> Compatibility and WebView Application Risks
>>>
>>> FLEDGE does not deprecate or change existing web behavior, so there
>>> should be no compatibility risk.
>>>
>>> Activation
>>>
>>> Successful testing of FLEDGE in-browser ad auctions requires
>>> participation from both parties selling ad space and advertisers buying ad
>>> space.  This level of cooperation during an origin trial is a heavy lift,
>>> but one that seems feasible given the high level of interest we’ve seen
>>> from ad techs.
>>>
>>> Security
>>>
>>> FLEDGE involves downloading and running JavaScript functions, referred
>>> to as worklets.  Chrome runs worklets from different origins in separate
>>> processes in very constrained environments to limit security
>>> vulnerabilities.  You can read more about these constraints and security
>>> considerations here
>>> <https://github.com/WICG/turtledove/blob/main/Original-TURTLEDOVE.md#security-considerations>
>>> .
>>>
>>> Privacy
>>>
>>> Unlike third-party cookies which are readable across sites, FLEDGE
>>> intends to keep interest group information from being exposed to sites. For
>>> example this is why there is no navigator.getAdInterestGroups() API. FLEDGE
>>> worklets, which can read individual interest groups, are isolated and
>>> cannot access the network, access storage or postMessage() to other
>>> contexts. As the proposed first FLEDGE origin trial details document
>>> <https://github.com/WICG/turtledove/blob/main/Proposed_First_FLEDGE_OT_Details.md>
>>> discusses, this first origin trial will not initially enable all of the
>>> isolation and privacy controls in order to ease developer testing. Over
>>> time, these privacy protections will be added as we introduce new releases.
>>> To protect user privacy and honor user choice, we will not enable this
>>> first origin trial for users that have disabled 3rd party cookies. We
>>> anticipate that this will also help ad-techs to more realistically compare
>>> their existing ad selection methods relying on third-party cookie
>>> availability.  Privacy considerations for the overall TURTLEDOVE family are
>>> discussed here
>>> <https://github.com/WICG/turtledove/blob/main/Original-TURTLEDOVE.md#privacy-considerations>
>>> .
>>>
>>> Browser Performance
>>>
>>> If advertisers place users into large numbers of interest groups, there
>>> is a risk of on-device FLEDGE ad auctions consuming excessive amounts of
>>> processing resources which could make the auctions introduce significant
>>> latency into the ad serving process, or slowing down the overall browsing
>>> experience.  This is an area of intense discussion, investigation, and
>>> improvement.
>>>
>>> Goals for experimentation
>>>
>>> Shifting interest data and final ad decision browser-side instead of
>>> server-side represents a major shift in interest group based advertising.
>>> We hope to get feedback from ad tech on FLEDGE’s effectiveness and
>>> performance.
>>>
>>> Experiment Configuration
>>>
>>> The origin trial for this experiment will be shared among various
>>> Privacy Sandbox APIs. Our goal is to allow for coordinated experiments to
>>> be run by multiple different sites, across multiple APIs in one OT.
>>>
>>> This shared origin trial, Privacy Sandbox Ads APIs, will be a
>>> third-party origin trial. To ensure that developers can run coordinated
>>> experiments without concern for exceeding page load usage thresholds, this
>>> Origin Trial will be available for a subset of users by default. Therefore,
>>> it will be necessary to feature test to ensure that the API surface you
>>> want to use is available after providing your OT token. A second advantage
>>> of this configuration is that different experimenters will experiment with
>>> the same users, which enables coordination for APIs like FLEDGE across
>>> third parties.
>>>
>>> Ongoing technical constraints
>>>
>>> FLEDGE depends on several other in-development web technologies, e.g.
>>> Fenced Frames, trusted key-value servers, and aggregate reporting.  To ease
>>> developer testing and measurement, this first FLEDGE origin trial will not
>>> require use of these other in-development web technologies.  For details of
>>> exactly what we’re proposing including in this first FLEDGE origin trial
>>> and why please read
>>> https://github.com/WICG/turtledove/blob/main/Proposed_First_FLEDGE_OT_Details.md
>>>
>>> Debuggability
>>>
>>> FLEDGE worklets can be debugged in Chrome’s Developer tools, instructions
>>> here
>>> <https://developer.chrome.com/blog/fledge-api/#debug-fledge-worklets>.
>>>
>>> FLEDGE interest groups can also be viewed in Chrome’s Developer tools:
>>> in the "Application" tab, there is an "Interest Groups" item on the left
>>> side-bar that, when clicked, should display all interest groups that this
>>> page interacted with, e.g. when a page joins/leaves an interest group, bids
>>> on an auction, or wins an auction on this page then the interest group
>>> should show up.
>>>
>>> Will this feature be supported on all six Blink platforms (Windows, Mac,
>>> Linux, Chrome OS, Android, and Android WebView)?
>>>
>>> No, this origin trial will be supported on all platforms except Android
>>> for reasons discussed here
>>> <https://github.com/WICG/turtledove/blob/main/Proposed_First_FLEDGE_OT_Details.md#mobile-devices>
>>> .
>>>
>>> Is this feature fully tested by web-platform-tests
>>> <https://chromium.googlesource.com/chromium/src/+/master/docs/testing/web_platform_tests.md>
>>> ?
>>>
>>> No. More web-platform-test
>>> <https://chromium.googlesource.com/chromium/src/+/master/docs/testing/web_platform_tests.md>
>>> coverage is expected when the specification is closer to completion.
>>>
>>> Flag name
>>>
>>> privacy-sandbox-ads-apis
>>>
>>> Requires code in //chrome?
>>>
>>> Nearly all code is outside //chrome, the exception being the related
>>> Privacy Sandbox Settings UI.
>>>
>>> Launch bug
>>>
>>> https://bugs.chromium.org/p/chromium/issues/detail?id=1181739
>>>
>>> Estimated milestones
>>>
>>> We hope to start the Origin Trial sometime during M101 beta. We plan to
>>> continue the Origin Trial until at least M104 to give developers time to
>>> test the API and provide feedback. Once we are confident that the APIs are
>>> working properly, we will transition the OT from beta to stable channel.
>>>
>>
>> Same question as on
>> https://groups.google.com/a/chromium.org/g/blink-dev/c/jEnNpideO1Y/m/5gSCiXUtAQAJ
>> regarding OT duration.
>>
>
> We're planning to start in M101 beta. We plan to progress to stable in
> M101 or later, if issues arise, and to continue the origin trial through
> M104, inclusive.
>
>
>>
>>
>>> Link to entry on the Chrome Platform Status
>>>
>>> https://chromestatus.com/feature/5733583115255808
>>>
>>> Links to previous Intent discussionsIntent to prototype:
>>> https://groups.google.com/a/chromium.org/g/blink-dev/c/w9hm8eQCmNI
>>>
>> --
> You received this message because you are subscribed to the Google Groups
> "blink-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to blink-dev+unsubscr...@chromium.org.
> To view this discussion on the web visit
> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CABQTWrmTRunapDYR4hMfmFyv2Cp5OpXTwNBqOLJJXdmLrn%3Dp6g%40mail.gmail.com
> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CABQTWrmTRunapDYR4hMfmFyv2Cp5OpXTwNBqOLJJXdmLrn%3Dp6g%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>

-- 
You received this message because you are subscribed to the Google Groups 
"blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to blink-dev+unsubscr...@chromium.org.
To view this discussion on the web visit 
https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAL5BFfVFe5kgHudidComkMP3wZeVNTsDpiXf9Uwr7W-RyRvCUQ%40mail.gmail.com.

Reply via email to