DNS-Based Service Discovery (DNS-SD) Privacy and Security Requirements
RFC 8882

Note: This ballot was opened for revision 05 and is now closed.

Éric Vyncke Yes

Alvaro Retana No Objection

Benjamin Kaduk No Objection

Comment (2020-03-04 for -05)
Section 1

   connected to the same network.  Consider for example a traveler
   wanting to upload pictures from a phone to a laptop when connected to
   the Wi-Fi network of an Internet cafe, or two travelers who want to

[both devices are on the same Wi-Fi, right?]

   Disclosing Information  In this document "disclosing information" is
      also focused on disclosure by data conveyed via messages on the
      service discovery protocol layer.

This is generic non-identity but still potentially sensitive data,
right?

Section 3.2

   kinds of means for making DNS-SD resource records available.  These
   means comprise but are not limited to mDNS [RFC6762], DNS servers
   ([RFC1033] [RFC1034], [RFC1035]), e.g. using SRP
   [I-D.ietf-dnssd-srp], and multi-link [RFC7558] networks.

nit: this "e.g." seems out of place.

Section 3.2.2

There is, of course, also no authentication requirement to claim a
particular instance name, so an active attacker can provide resources
that claim to be Alice's but are not.

Section 3.3.2

This sort of problem frequently ends up with a third-party "trusted
introducer", though it's not clear that mentioning this in the document
will add value.

3.4.2

I'm given to understand that for many radio technologies, multicast is
both effectively broadcast and has specific spectrum
requirements/properties that make it especially scarce, compared to
unicast spectrum.

Roman Danyliw No Objection

Comment (2020-03-04 for -05)
Thanks for this document.  Kuddos on the amazing ASCII art.

** Section 3.1.1. Per “Identifying devices leads to identifying people, either just for tracking people or as a preliminary to targeted attacks.”, this didn’t parse for me and the intent of the “just for tracking people” wasn’t clear.  Is the following the intent:

“Identifying devices can lead to identifying people, either for surveillance of these individuals in the physical world or as a preliminary step for a targeted cyber attack.”

** Section 3.1.2.  Per “The requirement in that scenario is that the discovery activity should not disclose the identify of either the client or the server”, is something stronger more desirable?  For example, is there any desire to thwart the discovery of the “business and social interactions” between the device owners?

** Section 3.1.3  It seems as if all of the same challenges of Section 3.1.1 “identifying people” and using the information for a “targeted attack” apply here too (but it’s said in a different way).  Is it worth link the same issues across scenarios?

** Section 3.2.  Per “Information conveyed via multicast messages can be obtained by an on-link attacker, while unicast messages are only available to MITM attackers.”, please clarify why a passive on-link attacker can’t see the unicast messages?

** Section 3.2.5.  Per “This combination of services and attributes will often be sufficient to identify the version of the software running on a device”, makes sense.  Is it worth adding that with this information and traffic analysis, you might also be able to get identity (track people).

(Alexey Melnikov; former steering group member) Yes

Yes ( for -05)
No email
send info

(Alissa Cooper; former steering group member) Yes

Yes (2020-03-04 for -05)
I was missing our old friends Alice and Bob in 3.1, but then saw that they return in later sections. It might be good to name the example users consistently.

(Barry Leiba; former steering group member) Yes

Yes (2020-02-25 for -05)
It’s useful to have this analysis; thanks.  Just some editorial comments below.  Please consider them; none needs any explicit response.  Please take specific note of the last one, about the references.

General: “i.e.” and “e.g.” always need a comma after them.

— Section 1 —

   There are cases when nodes connected to a network want to provide or
   consume services without exposing their identity to the other parties

Nit: “their identities” (or “a node… wants… its identity”)

   Consider for example a traveler

Nit: “Consider, for example, a traveler”

   Disclosing Information  In this document "disclosing information" is
      also focused on disclosure by data conveyed via messages on the
      service discovery protocol layer.

Do you mean “disclosure of data” (not “by”)?

— Section 2 —

   All these attackers can either be passive, i.e.
   they just listen to network traffic they have access to, or active,
   i.e. they additionally can craft and send (malicious) packets.

Style: You decide, of course, but I find this easier to read with parentheses, rather than “i.e.”s:

SUGGEST
   All these attackers can either be passive (they
   just listen to network traffic they have access to) or active
   (they additionally can craft and send malicious packets).
END

   on-link  An on-link attacker is on the same network link as victim
      devices engaging in service discovery; thus, the external attacker
      is in the same multicast domain.

The second line should say “on-link attacker”.

   MITM  A Man in the Middle (MITM) attacker either controls (parts of)
      a network link or can trick two parties to send traffic via him;

Nit: “Man-in-the-Middle” needs hyphens when it modifies “attacker”.
Style: I know that “him” matches “Man”, so maybe we should leave it as is.  Still, it jarred me.  I would say “via the attacker.”

— Section 3.1.1 —
I love the ASCII-art stick figures.  :-)

   just for tracking people or as a preliminary to targeted attacks.

“preliminary” isn’t a noun.  Maybe “preliminary step”, or maybe “preamble”?  Or you could remove “as a”, and it would work.  Yes, I think the last is the best fix here.

— Section 3.1.2 —

   such as
   for example an airport's lounge.

Nit: “for example” needs to be set off with commas before and after it.

— Section 3.1.3 —

   to further attacks, from theft to device specific hacking.

Nit: hyphenate “device-specific”.

   "fingerprint" of the person, allowing identification.

Style: I would say “facilitating identification”, or maybe “risking identification”.

— Section 3.2 —

   This is mainly relevant for unicast based discovery

Nit: hyphenate “unicast-based”.

— Section 3.2.4 —

   o  Some attributes like the paper size available in a printer,

Fruit flies like a banana?  The attributes are not paticularly fond of anything: “Some attributes, such as the paper size available in a printer,”

   Combinations of attributes have more information power than specific
   attributes

Style: I would say, “than individual attributes”

   Information contained in TXT records does not only breach privacy

Nit: make it “…records not only breaches privacy”

   Further, TXT records often contain version information about services
   allowing potential attackers

You need a comma after “services” — otherwise, the meaning isn’t quite as you want it.

— Section 3.2.5 —

   An argument is sometimes made that devices providing services can be
   identified by observing the local traffic, and that trying to hide
   the presence of the service is futile.  However,

Given what you say below this, I think it would help to emphasize the point here, so may I suggest this?:

NEW
   An argument is sometimes made that devices providing services can be
   identified by observing the local traffic, and that trying to hide
   the presence of the service is futile.  However, there are good
   reasons for the doscovery service layer to avoid unnecessary exposure:
END

— Section 3.2.6 —

   services, such as for example some private messaging services.

“such as” already means “for example”, so you don’t need both.  I would just remove “for example”.

— Section 3.4.1 —

   can perform, the proxy may have some way to wake the device, as
   implied in RFC6762 [RFC6762]

6762 is 50 pages-ish; do you mind adding a section to the citation to help the reader find the implication?

— Section 3.4.2 —

   Further it may cause an unnecessary level
   of power consumption which is particularly problematic

Nit: this needs a comma after “further” and another after “consumption”.

   unauthorized observers, while also managing to do that efficiently.

You’re missing an antecedent to “that”.  I think you need to say, “to do its job efficiently,” or something like that.

— Section 3.4.3 —

   establishment requires active participation of the user, such as
   entering a password or PIN.

I submit that “clicking OK” is also active participation.  Maybe “more significant participation” is better?

— Section 4 —

   lead to a solution that does not transmit privacy violating DNS-SD

Nit: hyphenate “privacy-violating”.

   mechanisms should be used on deeper layer network protocols and on
   how to actually connect to services in a privacy preserving way,

Nit: hyphenate “deeper-layer” and “privacy-preserving”.

— Section 4.2 —

   4.  Avoid disclosure of information about the services they offer to
       unauthorized clients.

This sounds like it’s talking about services that they offer to unauthorized clients.  I don’t actually think readers will misunderstand it, but they might trip over it.  Maybe move “to unauthorized clients” after “disclosure”?  That way, you can make the same change in bullet 3 and keep them parallel and clear.

— Section 4.3 —

       Further, it would
       increase power consumption which is critical for IoT devices.

Increased power consumption isn’t what’s critical; just the opposite.  Maybe “which is damaging to IoT devices.”

— Section 7 —
Do with this comment what you will: I’m one who believes that Informational documents do have Normative References.  Those are the references that are critical to the understanding of the document.  Clearly, the DNSSD and mDNS documents are in that category, and I think there are others.  You needn’t reply on this, and you needn’t do it if you disagree, but I think it would be best to identify the key documents that readers of this need to be familiar with, and move them into Normative References.

(Suresh Krishnan; former steering group member) Yes

Yes ( for -05)
No email
send info

(Adam Roach; former steering group member) No Objection

No Objection (2020-03-04 for -05)
Section 3.2:

>   Information conveyed via multicast messages can be
>   obtained by an on-link attacker, while unicast messages are only
>   available to MITM attackers.

I don’t think this is accurate. Given that many of the environments under consideration (e.g., airport WiFi) use unencrypted wireless transmission combined with a captive portal. In these cases, an eavesdropper on the same channel can snoop on even unicast traffic without mounting an MITM attack.

----------------------------------------------------------------------

General:

The document speaks of randomization of identifiers, including those commonly used by users to identify which services they want to connect to. While the current state of affairs may list a directory such as:

•	Adam’s iPhone
•	David’s Google Pixel 3
•	Alice’s Laptop

(allowing me to select something based on its published name)

This document seems to propose a future state where such directories are instead presented as:

•	{da566203-0320-4604-aa14-f58ae7bea00c}
•	{6c0952a5-a573-4d92-9d4a-a4bc111a35d8}
•	{785bed6b-1355-4e7e-ad57-b5ce27e83e56}

I find it a bit surprising that this document doesn’t include at least a cursory mention of the difficulty users may have in device rendezvous under such a scheme and potential solutions to such issues (e.g., using RFID or QR codes to provide pairing information).

(Deborah Brungard; former steering group member) No Objection

No Objection ( for -05)
No email
send info

(Mirja Kühlewind; former steering group member) No Objection

No Objection (2020-03-02 for -05)
nit sec 2:
"on-link  An on-link attacker is [...]; thus, the external attacker
      is in the same multicast domain."
s/external attacker/on-link attacker/

I agree with Barry that at least RFC6763 and RFC7558 should be normative references.

Also thanks for the quick reply to the TSV-ART review (and thanks Tommy for the review!)!