User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Firefox/60.0

Steps to reproduce:

Firefox supports DoH and there is an enhancement request for DoT.
https://bugzilla.mozilla.org/show_bug.cgi?id=1434852
https://bugzilla.mozilla.org/show_bug.cgi?id=1348406

With the availability of encrypted DNS transports in firefox traffic analysis mitigations like padding are becoming relevant.

This enhancement request is about adding support for EDNS padding support to firefox.

related RFCs:

https://tools.ietf.org/html/rfc8310#section-11.1
https://tools.ietf.org/html/rfc7830
https://tools.ietf.org/html/rfc8467

Is this a duplication of bug 1542754, nusenu?

You linked to bug 1542754 which is about supporting ESNI without requiring DoH, I don't see any overlap there.

one more note:

padding could also happen on the HTTP/2 layer as mentioned in
https://tools.ietf.org/html/rfc8484#section-4.1
https://tools.ietf.org/html/rfc7540#section-10.7

I don't think that either RFC 8484 or RFC 7540 define the specifics of the proposed padding.

Moreover, if the goal is to find a good balance between privacy and performance (i.e., don't pad the link to capacity) then the padding should probably be tailored to the respective application, i.e., DNS.

RFC 8467 (mentioned in the description of this bug) seems to describe a specific and reasonable padding strategy.
Specifically section 4.1 "Block-Level Padding" produces query sizes that are multiples of 128 bytes.
Response sizes are multiples of 468 bytes.

Someone could for instance study the DNS queries and responses generated when visiting the top 1,000 websites according to Cisco Umbrella (http://s3-us-west-1.amazonaws.com/umbrella-static/index.html) and get an idea of out how much the data overhead would be.

Depends on D126480

Depends on D126953

https://hg.mozilla.org/mozilla-central/rev/5b0d115a9b6e
https://hg.mozilla.org/mozilla-central/rev/cfdd0fddbce9
https://hg.mozilla.org/mozilla-central/rev/6ee09543a37f