[Experiment] Pref-Flip Experiment: Investigating network tolerance for new TLS extensions Fx 105.0 to 106.0 Beta
Categories
(Shield :: Shield Study, task, P3)
Tracking
(Not tracked)
People
(Reporter: experimenter, Assigned: djackson)
References
()
Details
User Story
Delivery Type: Pref Flip Experiment
What is the preference we will be changing
security.tls.ech.grease_probability
What are the branches of the experiment and what values should
each branch be set to?
- Treatment GREASE Probability 20% 50%:
Value: 20
This branch will use GREASE on roughly 20% of their connections.
- Control GREASE Disabled 50%:
Value: 0
This branch will have GREASE disabled.
What version and channel do you intend to ship to?
25% of Beta Firefox 105.0 to 106.0
Are there specific criteria for participants?
Experiments:
Exclude clients enrolled in "Pref-Flip Experiment: ECH - US & CA Interoperability Experiment"
Countries: all
Locales: all
What is your intended go live date and how long will the experiment run?
Sep 12, 2022 - Oct 10, 2022 (28 days)
What is the main effect you are looking for and what data will you use to
make these decisions?
For hypothesis 1 we will monitor existing telemetry in Firefox such as: `PERF_FIRST_CONTENTFUL_PAINT_MS` and `HTTP_PAGE_TLS_HANDSHAKE`. We expect to see no differences in the distributions between control and experimental branches. We will also monitor the specific telemetry: `SSL_HANDSHAKE_RESULT_FIRST_TRY` and `SSL_HANDSHAKE_RESULT_ECH_GREASE` which record the success or failures of their respective TLS connections. Plus the timing information recorded in `SSL_TIME_UNTIL_READY_ECH_GREASE` and `SSL_TIME_UNTIL_READY_FIRST_TRY`.
Within the experimental branch, we will monitor `SSL_HANDSHAKE_RESULT_FIRST_TRY` and `SSL_HANDSHAKE_RESULT_ECH_GREASE` for individual sessions and calculate the appropriate success probabilities.
Further details are contained in the experiment Gdoc: https://docs.google.com/document/d/1d49HEiOrQdicEv_S5fG93iVYXBKNyNEfaa2_YIZGLe0/edit?usp=sharing
Who is the owner of the data analysis for this experiment?
None
Will this experiment require uplift?
False
QA Status of your code:
Not Started.
Link to more information about this experiment:
https://experimenter.services.mozilla.com/experiments/investigating-network-tolerance-for-new-tls-extensions/
Investigating network tolerance for new TLS extensions
TLS is a network protocol used by Firefox to communicate securely with websites. Previous attempts to migrate to new versions of TLS or enable new TLS extensions have caused network connections to fail due to bugs in TLS Server implementations or network middleboxes.
This study will enable a new GREASE extension in the TLS Client Hello which will be filled with random data but will be otherwise ignored by Firefox and correct server implementations. The study will measure how this addition impacts the probability of a TLS connection succeeding and quantify any impact on performance due to the first network packet being larger.
Experimenter is the source of truth for details and delivery. Changes to Bugzilla are not reflected in Experimenter and will not change delivery configuration.
Data Science Issue: https://mozilla-hub.atlassian.net/browse/DO-831
More information: https://experimenter.services.mozilla.com/experiments/investigating-network-tolerance-for-new-tls-extensions/
|
||
Comment 1•2 years ago
|
||
We have finished testing the experiment and sent a GREEN sign off. You can find it here.
|
Reporter | |
Updated•2 years ago
|
|
|
Reporter | |
Comment 2•2 years ago
|
||
|
|
Reporter | |
Updated•2 years ago
|
Description
•