Re: [blink-dev] Intent to Experiment: Network State Partitioning (once more, with feeling)

2022-10-05 Thread Mike West 
You've got your LGTM already, but I'll double-LGTM it just to say that I
appreciate the care you're putting into gathering data to support the
decisions we make here. Thank you.

-mike


On Wed, Oct 5, 2022 at 10:01 AM Yoav Weiss  wrote:

> LGTM for a 1% stable experiment
>
> On Wed, Oct 5, 2022 at 12:20 AM Mike Taylor 
> wrote:
>
>> *Contact emails*
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> * brgoldst...@chromium.org ,
>> miketa...@chromium.org , mme...@chromium.org
>>   Summary We would like to run another experiment for
>> the Network State Partitioning effort to better understand the performance
>> impacts of an alternative partitioning scheme. To date we’ve run
>> experiments with a triple-keyed HTTP cache and triple-keyed network state,
>> and a double-keyed HTTP cache and double-keyed network state.  As a
>> possible compromise between security and performance tradeoffs, which were
>> discussed
>> 
>> at a recent WebAppSec session at TPAC, we want to measure a triple-keyed
>> HTTP cache with the following network state key configurations: -
>> Triple-keyed network state (this is the first experiment we ran, but the
>> data is quite old now) - Double-keyed network state - Double-key with an
>> is_cross_site bit (whether the iframe is cross-site from the top-level
>> parent frame). - (and a control group, with unpartitioned network state) As
>> a reminder, here’s a link to the (triple-key) I2E
>> ,
>> and the (double-key) I2E
>> .
>> This document only contains changes to previous intents. Blink component
>> Internals>Network
>> 
>> Goals for experimentation We want to roll this out to 1% of Stable traffic
>> (i.e., this is not an Origin Trial) to be able to compare to the previous
>> experiments. Estimated milestones We’d like to run at 1% on Stable for ~4
>> weeks, ideally beginning in M107 (but some chance of this slipping to M108
>> depending on timing). Tracking bug https://crbug.com/993801
>>  Launch bug https://crbug.com/1166303
>>  Link to entry on the Chrome Platform Status
>> https://chromestatus.com/feature/6713488334389248
>>  *
>>
>> --
>> 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/2e1e2e27-514b-688b-6318-e029fa7bf08f%40chromium.org
>> 
>> .
>>
> --
> 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/CAL5BFfXWnY4jM9DDy9X3j75xafDA%3DQ79mQSnA6A4GRiDopP1Zg%40mail.gmail.com
> 
> .
>

-- 
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/CAKXHy%3DcFBC5exY6uHEunS0Nb_pQ5S8pu_Xx4w4iqVcgMB2pmcA%40mail.gmail.com.

Re: [blink-dev] Intent to Experiment: Network State Partitioning (once more, with feeling)

2022-10-05 Thread Yoav Weiss 
LGTM for a 1% stable experiment

On Wed, Oct 5, 2022 at 12:20 AM Mike Taylor  wrote:

> *Contact emails*
>
>
>
>
>
>
>
>
>
>
>
>
>
> * brgoldst...@chromium.org ,
> miketa...@chromium.org , mme...@chromium.org
>   Summary We would like to run another experiment for
> the Network State Partitioning effort to better understand the performance
> impacts of an alternative partitioning scheme. To date we’ve run
> experiments with a triple-keyed HTTP cache and triple-keyed network state,
> and a double-keyed HTTP cache and double-keyed network state.  As a
> possible compromise between security and performance tradeoffs, which were
> discussed
> 
> at a recent WebAppSec session at TPAC, we want to measure a triple-keyed
> HTTP cache with the following network state key configurations: -
> Triple-keyed network state (this is the first experiment we ran, but the
> data is quite old now) - Double-keyed network state - Double-key with an
> is_cross_site bit (whether the iframe is cross-site from the top-level
> parent frame). - (and a control group, with unpartitioned network state) As
> a reminder, here’s a link to the (triple-key) I2E
> ,
> and the (double-key) I2E
> .
> This document only contains changes to previous intents. Blink component
> Internals>Network
> 
> Goals for experimentation We want to roll this out to 1% of Stable traffic
> (i.e., this is not an Origin Trial) to be able to compare to the previous
> experiments. Estimated milestones We’d like to run at 1% on Stable for ~4
> weeks, ideally beginning in M107 (but some chance of this slipping to M108
> depending on timing). Tracking bug https://crbug.com/993801
>  Launch bug https://crbug.com/1166303
>  Link to entry on the Chrome Platform Status
> https://chromestatus.com/feature/6713488334389248
>  *
>
> --
> 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/2e1e2e27-514b-688b-6318-e029fa7bf08f%40chromium.org
> 
> .
>

-- 
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/CAL5BFfXWnY4jM9DDy9X3j75xafDA%3DQ79mQSnA6A4GRiDopP1Zg%40mail.gmail.com.

[blink-dev] Intent to Experiment: Network State Partitioning (once more, with feeling)

2022-10-04 Thread Mike Taylor 

**


   *Contact emails*

*

brgoldst...@chromium.org , 
miketa...@chromium.org , 
mme...@chromium.org 



   Summary

We would like to run another experiment for the Network State 
Partitioning effort to better understand the performance impacts of an 
alternative partitioning scheme. To date we’ve run experiments with a 
triple-keyed HTTP cache and triple-keyed network state, and a 
double-keyed HTTP cache and double-keyed network state.



As a possible compromise between security and performance tradeoffs, 
which were 
discussed 
at 
a recent WebAppSec session at TPAC, we want to measure a triple-keyed 
HTTP cache with the following network state key configurations:


 *

   Triple-keyed network state (this is the first experiment we ran, but
   the data is quite old now)

 *

   Double-keyed network state

 *

   Double-key with an is_cross_site bit (whether the iframe is
   cross-site from the top-level parent frame).

 *

   (and a control group, with unpartitioned network state)


As a reminder, here’s a link to the (triple-key) I2E 
, 
and the (double-key) I2E 
. 
This document only contains changes to previous intents.



   Blink component

Internals>Network 




   Goals for experimentation

We want to roll this out to 1% of Stable traffic (i.e., this is not an 
Origin Trial) to be able to compare to the previous experiments.



   Estimated milestones

We’d like to run at 1% on Stable for ~4 weeks, ideally beginning in M107 
(but some chance of this slipping to M108 depending on timing).



   Tracking bug

https://crbug.com/993801 


   Launch bug

https://crbug.com/1166303 



   Link to entry on the Chrome Platform Status

https://chromestatus.com/feature/6713488334389248 



*

--
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/2e1e2e27-514b-688b-6318-e029fa7bf08f%40chromium.org.