Fwd: Review of: draft-ietf-v6ops-v6-aaaa-whitelisting-implications-03 *(formal for apps area)*

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 5/29/2011 7:46 AM, Livingood, Jason wrote:
> Hi Bernard – I've finally found the time to close out the last bits of
> feedback > in this version of the draft.


Jason,

Perhaps my filters misfiled your followup to the "formal" review I was asked to do, or perhaps my earlier, informal, and vary narrow review muddied the waters, but I am not finding your response to the Apps Area review.

For convenience, here it is again.

d/

-------- Original Message --------
Subject: Review of:  draft-ietf-v6ops-v6-aaaa-whitelisting-implications-03	
         *(formal for apps area)*
Date: Mon, 09 May 2011 10:18:51 -0700
From: Dave CROCKER <dhc@xxxxxxxxxxxx>
Reply-To: dcrocker@xxxxxxxx
Organization: Brandenburg InternetWorking
To: IETF Discussion <ietf@xxxxxxxx>
CC: v6ops@xxxxxxxx <v6ops@xxxxxxxx>, Apps Review <apps-review@xxxxxxxx>

(This is an "official" and significantly extended version of an informal and
narrow review I posted earlier.  /d)



Howdy.

I have been selected as the Applications Area Review Team reviewer for this
draft (for background on apps-review, please see
http://www.apps.ietf.org/content/applications-area-review-team).

Please resolve these comments along with any other Last Call comments you may
receive. Please wait for direction from your document shepherd or AD before
posting a new version of the draft.



Review (v2):

Title:  IPv6 AAAA DNS Whitelisting Implications
I-D:    draft-ietf-v6ops-v6-aaaa-whitelisting-implications-03

By:     D. Crocker <dcrocker@xxxxxxxx>
Date:   <>



Summary
=======

This draft covers a a dual-stack problem in which a target host's DNS entry
contains records for IPv4 and IPv6, but returning IPv6 information to a DNS
client can cause problems. The paper discusses for resolving this through use of
a a DNS-based mechanism that manually lists response preferences to select which
DNS records to return.  The paper describes the mechanism and explores various
effects and possibilities of its use, including the difference between using it
selectively among a smaller number of sites, versus universally.

The draft is a serious effort to explore the use of such a mechanism and it
touches many different issues.  It is generally well-organized and clearly
written, although it very much needs the aid of a professional technical editor.
The writing often assumes too much knowledge by the reader.

The paper's exploration of universal adoption seems to vary between considering
that goal practical versus considering it only as a matter of completeness for
discussing the full range of possibilities.  That is, it is not clear whether
the paper views this alternative as practically possible and even preferred,
versus only a matter for academic thoroughness. The paper needs to take a basic
position about feasibility, explain it in terms of comparable adoption efforts
at Internet-scale, and then make its treatment of universal adoption a bit more
consistent.

When introducing terms, mechanisms, configurations and scenarios, the paper
needs to be more careful to describe them adequately for a reader new to the
topic.  This is not a matter of having a tutorial about the DNS, but rather a
tutorial for this type of mechanism and when and how it can be used.

As a specific example, the document cites "domain-by-domain" use, but I am not
clear how that would work, in terms of configuration and cross-net information
exchange.  One question is how the server knows the 'domain' of the client?

The document should careful to distinguish what is existing practice, versus
what is being explored as added possibilities.  The difference in concreteness
and certitude between the two is substantial.

The document's use of the term whitelisting appears to continue an existing,
recent use, for this type of mechanism.  Unfortunately it directly conflicts
with long-standing use of the term by the anti-abuse community for whitelisting
in the DNS. Its use here also seems to be a mismatch with the word's dictionary
semantics, which is most naturally used to distinguish yes/no choices, rather
than either/or choices.  So there is no intuitive sense of "goodness" (whitelist
= yes) or "badness" (blacklist) for this use. The word "preferences" seems more
in line with the meaning of the mechanism.



Detailed Comments
=================


Abstract

   The objective of this document is to describe what the whitelisting
   of DNS AAAA resource records is, hereafter referred to as DNS

RRs are whitelisted?  Isn't it the addresses and not the records that are
whitelisted?

Does this mean putting whitelisting records into the DNS or does it mean
something else?

Comcast's own considerable expertise notwithstanding, has this doc been vetted
with a range of organizations that actually DO whitelisting?  Has it been
circulated through MAAWG and APWG?  Any comments from Spamhaus?  The
Acknowledgements list does not seem to indicate a range of whitelist ops folks
whose names I know.  (But then, I only know a few...)


   whitelisting, as well as the implications of this emerging practice
   and what alternatives may exist.  The audience for this document is
   the Internet community generally, including the IETF and IPv6
   implementers.

I suspect that product marketers won't have much interest in this.  I suspect
that the target for this is anti-abuse technical and operations staff. In any
event, the targetting statement should be more precise.


1.  Introduction

   This document describes the emerging practice of whitelisting of DNS

One natural, semantic problem with the term 'whitelist' is that it does not
really match the function being performed.  The white/black distinction implies
goodness -- or as Wikipedia says, "priviledge".  Instead, the use here is for
preference or priority.  What would a "blacklist" be, here?  Also note it is not
obvious what it means to be whitelisted, here?  Does it mean to choose the AAAA
records or the A records?

This is more like a 'Preference' or 'Configuration' list.

At the least, the name for this should be IPv6 Resolver Whitelisting.  It makes
clear /what/ is being "whitelisted".


   AAAA resource records (RRs), which contain IPv6 addresses, hereafter
   referred to as DNS whitelisting.  The document explores the

This provides a name, but not a function.  That is, it does not say what this
mechanisms actually /does/ or is /for/.


   implications of this emerging practice are and what alternatives may
   exist.

   The practice of DNS whitelisting appears to have first been used by
   major web content sites (sometimes described herein as "highly-

It's use for email anti-abuse dates back farther.

   <http://www.dnswl.org/>

   <http://en.wikipedia.org/wiki/DNSBL>


<http://publib.boulder.ibm.com/infocenter/domhelp/v8r0/index.jsp?topic=/com.ibm.help.domino.admin.doc/DOC/H_USING_DNS_whitelists_OVER.html>

Specifically within the context of the DNS, the term whitelisting is therefore
made ambiguous.

A google query for "whitelist dns" also demonstrates the history and current
ambiguity.


   trafficked domains" or "major domains").  These web site operators,
   or domain operators, observed that when they added AAAA resource
   records to their authoritative DNS servers in order to support IPv6
   access to their content that a small fraction of end users had slow
   or otherwise impaired access to a given web site with both AAAA and A
   resource records.  The fraction of users with such impaired access
   has been estimated to be roughly 0.078% of total Internet users
   [IETF-77-DNSOP] [NW-Article-DNSOP] [Evaluating IPv6 Adoption] [IPv6
   Brokenness].  Thus, in an example Internet Service Provider (ISP)
   network of 10 million users, approximately 7,800 of those users may
   experience such impaired access.

At a minimum, these sorts of statistics need to be normalized across IPv6
users/traffic, given how small a percentage that is, in total users and total
traffic.  If that's what is meant it should be stated.  If it isn't, the
statistic should be recalculated and explained a bit more precisely.


   As a result of this impairment affecting end users of a given domain,
   a few major domains have either implemented DNS whitelisting or are
   considering doing so [NW-Article-DNS-WL] [IPv6 Whitelist Operations].
   When implemented, DNS whitelisting in practice means that a domain's
   authoritative DNS will return a AAAA resource record to DNS recursive
   resolvers [RFC1035] on the whitelist, while returning no AAAA
   resource records to DNS resolvers which are not on the whitelist.  It

This explanation of the function should be offered sooner and should be
summarized in the Abstract.


   is important to note that these major domains are motivated by a
   desire to maintain a high-quality user experience for all of their

Rather than being important to note, this sentence sounds oddly like marketing
hype, in a technical specification.  It is gratuitous because specified features
are never added to /lower/ the quality of the user experience, for example.

In addition, the mechanism also affects client activity that has no user
directly involved.


   users.  By engaging in DNS whitelisting, they are attempting to
   shield users with impaired access from the symptoms of those
   impairments.

The /technical/ statement that should be here is that they are attempting to
provide a work-around for problematic behaviors in dual-stack IPv4/IPv6
environments.

The paper should make more clear exactly where the problem lies and when. If it
can occur for a number of reasons, explaining each of those scenarios would be
useful.


   Critics of the practice of DNS whitelisting have articulated several
   concerns.  Among these are that:

   o  DNS whitelisting is a very different behavior from the current
      practice concerning the publishing of IPv4 address resource
      records,

   o  that it may create a two-tiered Internet,

   o  that policies concerning whitelisting and de-whitelisting are
      opaque,





Livingood                Expires August 26, 2011                [Page 5]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   o  that DNS whitelisting reduces interest in the deployment of IPv6,

   o  that new operational and management burdens are created,

well, yeah... in fact it should be noted that the burdens are particularly
onerous at scale.


   o  and that the costs and negative implications of DNS whitelisting
      outweigh the perceived benefits, compared to fixing underlying
      impairments.

and it doesn't scale.

and it violates an extremely basic premise of cross-Internet interoperability by
requiring prior arrangement.


   This document explores the reasons and motivations for DNS
   whitelisting.  It also explores the outlined concerns regarding this
   practice.  Readers will hopefully better understand what DNS
   whitelisting is, why some parties are implementing it, and what
   criticisms of the practice exist.


2.  How DNS Whitelisting Works

How IPv6 AAAA DNS Whitelisting Works.

(Anti-spam DNS Whitelisting works rather differently...)


   DNS whitelisting is implemented in authoritative DNS servers.  These
   servers implement IP address-based restrictions on AAAA query
   responses.  So far, DNS whitelisting has been primarily implemented
   by web server operators deploying IPv6-enabled services.  For a given

Really?  This is web-specific?  The same restrictions are not applied for other
applications?

So if the same client-side hosts attempt to contact the server for email or
xmpp, they won't get the same handling?


   operator of a website, such as www.example.com, the operator
   essentially applies an access control list (ACL) on the authoritative
   DNS servers for the domain example.com.  The ACL is populated with

An ACL usually is a yes/no mechanism.  Here, however, the mechanism is for
asserting a preference for IPv6 over IPv4.

That does not seem to match the definition of ACL that I'm used to, unless the
semantic is defined as denying IPv4 access to the listed clients.

The term ACL is particularly odd to use if the mechanism pertains to responses
rather than queries.


   the IPv4 and/or IPv6 addresses or prefix ranges of DNS recursive

Either address type can be listed?  So this really is a pure 'preferences'
mechanism?

Which settings count as whitelisting?  Do any count as blacklisting?



   resolvers on the Internet, which have been authorized to receive AAAA
   resource record responses.  These DNS recursive resolvers are
   operated by third parties, such as ISPs, universities, governments,
   businesses, and individual end users.  If a DNS recursive resolver IS
   NOT matched in the ACL, then AAAA resource records will NOT be sent
   in response to a query for a hostname in the example.com domain.

This configuration appears to ensure the maximum barrier to adoption for IPv6,
since it means that IPv6 will not work automatically.  It will only work for
hosts that are manually configured to receive responses with v6 records.

That's a rather major implication.  It's a default that is probably meant to
apply during the very early stages of adoption, when there are few users of the
newer mechanism.

It's probably worth discussing it in more detail, including discussing when to
change the default...


   However, if a DNS recursive resolver IS matched in the ACL, then AAAA
   resource records will be sent in response to a query for a given
   hostname in the example.com domain.  While these are not network-
   layer access controls they are nonetheless access controls that are a
   factor for end users and other parties like network operators,
   especially as networks and hosts transition from one network address
   family to another (IPv4 to IPv6).


Also, all of this clarifies the function of this listing mechanism and suggests
a very different name, to be more precise and accurate in naming it:

    IPv6 DNS Response Preference List.


   In practice, DNS whitelisting generally means that a very small
   fraction of the DNS recursive resolvers on the Internet (those in the
   whitelist ACL) will receive AAAA responses.  The large majority of
   DNS resolvers on the Internet will therefore receive only A resource
   records containing IPv4 addresses.  Thus, quite simply, the
   authoritative server hands out different answers depending upon who
   is asking; with IPv4 and IPv6 resource records for some on the
   authorized whitelist, and only IPv4 resource records for everyone
   else.  See Section 2.1 and Figure 1 for a description of how this



Livingood                Expires August 26, 2011                [Page 6]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   works.

   Finally, DNS whitelisting can be deployed in two primary ways:
   universally on a global basis, or on an ad hoc basis.  Deployment on
   a universal deployment basis means that DNS whitelisting is
   implemented on all authoritative DNS servers, across the entire
   Internet.  In contrast, deployment on an ad hoc basis means that only
   some authoritative DNS servers, and perhaps even only a few,
   implement DNS whitelisting.  These two potential deployment models
   are described in Section 6.

2.1.  Description of the Operation of DNS Whitelisting

   The system logic of DNS whitelisting is as follows:

   1.  The authoritative DNS server for example.com receives DNS queries
       for the A (IPv4) and AAAA (IPv6) address resource records for the
       FQDN www.example.com, for which AAAA (IPv6) resource records
       exist.

This means that the mechanism is /only/ triggered when /both/ address records
are queried?  A query for only one type of address record won't trigger the list
lookup?  I think that doesn't match other statements in the document.


   2.  The authoritative DNS server examines the IP address of the DNS
       recursive resolver sending the AAAA (IPv6) query.

"examines"?  Examines it for what?  What does this step mean?


   3.  The authoritative DNS server checks this IP address against the
       access control list (ACL) that is the DNS whitelist.

   4.  If the DNS recursive resolver's IP address IS matched in the ACL,
       then the response to that specific DNS recursive resolver can
       contain AAAA (IPv6) address resource records.

Oh.  This is not about whether to send responses /over/ v6 vs. v4?  This is
whether to /include/ a particular type of RR in responses???

In that case an appropriate name for this mechanism is more like:

   DNS Response Content Preference List

And this seems even less like an ACL than it did before.  (I assume the
justification is that access is being prevented by virtue of not supplying the
address, but still...)


   5.  If the DNS recursive resolver's IP address IS NOT matched in the
       ACL, then the response to that specific DNS recursive resolver
       cannot contain AAAA (IPv6) address resource records.  In this
       case, the server should return a response with the response code
       (RCODE) being set to 0 (No Error) with an empty answer section
       for the AAAA record query.















Livingood                Expires August 26, 2011                [Page 7]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   ---------------------------------------------------------------------
   A query is sent from a DNS recursive resolver that IS NOT on the DNS
   whitelist:

               Request                      Request
           www.example.com                  www.example.com
                 AAAA    +-------------+     AAAA    +-----------------+
     ++--++   ---------> |  RESOLVER   |  ---------> | www.example.com |
     ||  ||       A      | **IS NOT**  |      A      | IN A exists     |
   +-++--++-+ ---------> |     ON      |  ---------> | IN AAAA exists  |
   +--------+     A      | example.com |      A      |                 |
      Host    <--------- |  WHITELIST  |  <--------- |                 |
    Computer   A Record  +-------------+  A Record   +-----------------+
               Response   DNS Recursive   Response       example.com
              (only IPv4)   Resolver     (only IPv4)    Authoritative
                              #1                           Server
   ---------------------------------------------------------------------
   A query is sent from a DNS recursive resolver that IS on the DNS
   whitelist:

               Request                      Request
           www.example.com                  www.example.com
                AAAA     +-------------+     AAAA    +-----------------+
     ++--++   ---------> |  RESOLVER   |  ---------> | www.example.com |
     ||  ||       A      |   **IS**    |      A      | IN A exists     |
   +-++--++-+ ---------> |     ON      |  ---------> | IN AAAA exists  |
   +--------+   AAAA     | example.com |     AAAA    |                 |
      Host    <--------- |  WHITELIST  |  <--------- |                 |
    Computer      A      |             |      A      |                 |
              <--------- |             |  <--------- |                 |
              A and AAAA +-------------+ A and AAAA  +-----------------+
               Record     DNS Recursive   Record        example.com
              Responses     Resolver     Responses      Authoritative
              (IPv4+IPv6)      #2        (IPv4+IPv6)       Server
   ---------------------------------------------------------------------

              Figure 1: DNS Whitelisting - Functional Diagram

This diagram is confusing to me.  I suspect that a protocol exchange sequence
format, in the style of:

     Host             Resolver 1            Authoritative

          ---------->
                                 --------->
                                <---------
         <----------

will be considerably more helpful.


3.  What Problems Are Implementers Trying To Solve?

This is a very useful section and it is probably worth moving it higher, to
precede the 'how it works' section.


   As noted in Section 1, domains which implement DNS whitelisting are
   attempting to protect a few users of their domain, who have impaired
   IPv6 access, from having a negative experience (poor performance).

By the way, what does 'impaired v6 access' mean?

I think there needs to be a simple, direct description of what occurs without
this mechanism.

For example, perhaps you mean that a host can send DNS queries using IPv6 but
cannot receive DNS responses over IPv6? Perhaps you mean that the host can send
IPv6 but cannot receive it.  (That's a different scale and scope of problem from
the first example I gave.)

This brief, summary problem statement should be included in the Abstract, to
make /much/ more clear what this mechanism is for.


   While it is outside the scope of this document to explore the various
   reasons why a particular user's system (host) may have impaired IPv6
   access, for the users who experience this impairment it is a very
   real performance impact.  It would affect access to all or most dual



Livingood                Expires August 26, 2011                [Page 8]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   stack services to which the user attempts to connect.  This negative
   end user experience can range from someone slower than usual (as
   compared to native IPv4-based access), to extremely slow, to no
   access to the domain whatsoever.

Rather than repeat that this is about end-users, it sounds more that this is
about whether a service works or does not work, whether a user is directly
present or not.


   While one can debate whether DNS whitelisting is the optimal solution
   to the end user experience problem, it is quite clear that DNS
   whitelisting implementers are interested in maximizing the
   performance of their services for end users as a primary motivation
   for implementation.

You keep citing 'performance' but haven't described what sort of performance
degradation takes place. Is this really about relatively better or worse
performance -- and if so, how -- or is this about working or not working?

Also rather than saying what implementers are interested in, it's probably more
helpful to note that the practice is now significantly established and therefore
worth documenting, independent of its possible controversy.


   At least one highly-trafficked domain has noted that they have
   received requests to not send DNS responses with AAAA resource
   records to particular resolvers.  In this case, the operators of

"At least one" seems a rather tiny statistic.  Perhaps the actual statistic is
significantly larger?


   those recursive resolvers have expressed a concern that their IPv6

I suspect that it's not resolvers that are doing the expressing, since their
vocabulary is usually too limited...


   network infrastructure is not yet ready to handle the large traffic
   volume which may be associated with the hosts in their network
   connecting to the websites of these domains.  This concern is clearly

So even though the site allows v6 DNS queries to go out from a host, it can't
really support having the host use v6?

Wow. I do understand why service providers often have to work around silliness
at the client side, but this problem at the client side seems particularly
egregious.


   a temporary consideration relating to the deployment of IPv6 network
   infrastructure on the part of networks with end user hosts, rather
   than a long-term concern.  These end user networks may also have

Again this goal of short-term usage is worth noting earlier, including in the
Abstract.


   other tools at their disposal in order to address this concern,
   including applying rules to network equipment such as routers and
   firewalls (this will necessarily vary by the type of network, as well
   as the technologies used and the design of a given network), as well
   as configuration of their recursive resolvers (though modifying or
   suppressing AAAA resource records in a DNSSEC-signed domain on a
   Security-Aware Resolver will be problematic Section 10.1).

   Some implementers with highly-trafficked domains have explained that
   DNS whitelisting is a necessary, though temporary, risk reduction
   tactic intended to ease their transition to IPv6 and minimize any
   perceived risk in such a transition.  As a result, they perceive this
   as a tactic to enable them to incrementally enable IPv6 connectivity
   to their domains during the early phases of their transition to IPv6.

   Finally, some domains, have run IPv6 experiments whereby they added
   AAAA resource records and observed and measured errors [Heise Online
   Experiment], which should be important reading for any domain
   contemplating either the use of DNS whitelisting or simply adding
   IPv6 addressing to their site.


4.  Concerns Regarding DNS Whitelisting

   There are a number of potential implications relating to DNS
   whitelisting, which have been raised as concerns by some parts of the
   Internet community.  Many of those potential implications are further

I think the implications are not conditional; they exist rather than being
potential.  The 'potential' is that what is implicated will come to pass.




Livingood                Expires August 26, 2011                [Page 9]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   enumerated here and in Section 7.

Pro forma question:  Why are implications discussed in multiple places?


   Some parties in the Internet community, including ISPs, are concerned

This style of text personalizes the issues unnecessarily (IMO).  It does not
really matter who holds the concerns, or else they'd be described more precisely.

I suggest merely noting that there are concerns and then listing and discussing
the concerns, rather than adding text to attribute the concerns to others, even
if the conclusion of your text is that a particular concern is not valid.


   that the practice of DNS whitelisting for IPv6 address resource
   records represents a departure from the generally accepted practices
   regarding IPv4 address resource records in the DNS on the Internet
   [Whitelisting Concerns].  These parties explain their belief that for

"These parties explain their belief" is an example of personalization that is
not needed.  This isn't about the believers.  It is about possible problems.


   A resource records, containing IPv4 addresses, once an authoritative
   server operator adds the A record to the DNS, then any DNS recursive
   resolver on the Internet can receive that A record in response to a

This does not appear to be a grammatically valid sentence.  My guess is that
deleting "A resource... addresses" fixes this.

And by the way, the document's reference to "recursive" resolvers is mostly
likely incorrect.  The problem is not restricted only to that very specific type
of resolver, is it?

If in fact it /is/ specific to them -- and your following text describes an
indirect effects scenario where it might be -- I suggest calling out the
configuration at the beginning, along the lines of:

     One way the problem with returning AAAA records can be experienced is when
recursive resolvers are used.  Although that resolver might support IPv6, its
client hosts might not.  So, returning an AAAA record will mean that these
limited hosts will be given an unusable address.

And this type of description belongs in the text describing the motivating
problem(s), rather than buried in the 'concerns' discussion.

(The text, here, pertains to A records, but the problem I've described uses the
same configuration but for AAAA records with mixed v6 support.)


   query.  By extension, this means that any of the hosts connected to
   any of these DNS recursive resolvers can receive the IPv4 address
   resource records for a given FQDN.  This enables new server hosts
   which are connected to the Internet, and for which a fully qualified
   domain name (FQDN) such as www.example.com has been added to the DNS
   with an IPv4 address record, to be almost immediately reachable by
   any host on the Internet.  In this case, these new servers hosts
   become more and more widely accessible as new networks and new end
   user hosts connect to the Internet over time, capitalizing on and
   increasing so-called "network effects" (also called network
   externalities).  It also means that the new server hosts do not need
   to know about these new networks and new end user hosts in order to
   make their content and applications available to them, in essence
   that each end in this end-to-end model is responsible for connecting
   to the Internet and once they have done so they can connect to each
   other without additional impediments or middle networks or
   intervening networks or servers knowing about these end points and
   whether one is allowed to contact the other.

Hmmm.  This rather lengthy bit of prose appears merely to be explaining the
basic and long-standing DNS value proposition???


   In contrast, the concern is that DNS whitelisting may fundamentally
   change this model.  In the altered DNS whitelisting end-to-end model,
   one end (where the end user is located) cannot readily connect to the
   other end (where the content is located), without parts of the middle
   (recursive resolvers) used by one end (the client, or end user hosts)
   being known to an intermediary (authoritative nameservers) and
   approved for access to the resource at the end.  As new networks
   connect to the Internet over time, those networks need to contact any
   and all domains which have implemented DNS whitelisting in order to
   apply to be added to their DNS whitelist, in the hopes of making the
   content and applications residing on named server hosts in those
   domains accessible by the end user hosts on that new network.
   Furthermore, this same need to contact all domains implementing DNS
   whitelisting also applies to all pre-existing (but not whitelisted)
   networks connected to the Internet.

   In the current IPv4 Internet when a new server host is added to the
   Internet it is generally widely available to all end user hosts and
   networks, when DNS whitelisting of IPv6 resource records is used,

If it is available to the hosts, it is available to the network.

networks, when -> networks. When





Livingood                Expires August 26, 2011               [Page 10]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   these new server hosts are not accessible to any end user hosts or
   networks until such time as the operator of the authoritative DNS

They are still accessible.  The IP-level mechanisms still work.

They are not reachable when using the domain name.


   servers for those new server hosts expressly authorizes access to
   those new server hosts by adding DNS recursive resolvers around the
   Internet to the ACL.  This has the potential to be a significant

This is a good example of the reason the term ACL is inappropriate:  It implies
a security protection that does not actually exist.  The hosts are still accessible.


   change in reachability of content and applications by end users and
   networks as these end user hosts and networks transition to IPv6,
   resulting in more (but different) breakage.  A concern expressed is
   that if much of the content that end users are most interested in is
   not accessible as a result, then end users and/or networks may resist
   adoption of IPv6 or actively seek alternatives to it, such as using
   multi-layer network address translation (NAT) techniques like NAT444
   [I-D.shirasaki-nat444] on a long-term basis.  There is also concern
   that this practice also could disrupt the continued increase in
   Internet adoption by end users if they cannot simply access new
   content and applications but must instead contact the operator of
   their DNS recursive resolver, such as their ISP or another third
   party, to have their DNS recursive resolver authorized for access to
   the content or applications that interests them.  Meanwhile, these
   parties say, over 99.9% of the other end users that are also using
   that same network or DNS recursive resolver are unable to access the
   IPv6-based content, despite their experience being a positive one.

   While in Section 1 the level of IPv6-related impairment has been
   estimated to be as high as 0.078% of Internet users, which is a

8 hundredths of one percent?

That's considered a high percentage?

Even if it is 8%, is that considered high?


5.2.  Similarities to DNS Load Balancing

   DNS whitelisting also has some similarities to DNS load balancing.
   There are of course many ways that DNS load balancing can be
   performed.  In one example, multiple IP address resource records (A
   and/or AAAA) can be added to the DNS for a given FQDN.  This approach
   is referred to as DNS round robin [RFC1794].  DNS round robin may
   also be employed where SRV resource records are used [RFC2782].

Right, but that's algorithmic rather than involving the manual method, described
here. So it does not seem comparable.


6.  Likely Deployment Scenarios

   In considering how DNS whitelisting may emerge more widely, there are
   two likely deployment scenarios, which are explored below.

   In either of these deployment scenarios, it is possible that
   reputable third parties could create and maintain DNS whitelists, in
   much the same way that blacklists are used for reducing email spam.
   In the email context, a mail operator subscribes to one or more of
   these lists and as such the operational processes for additions and
   deletions to the list are managed by a third party.  A similar model
   could emerge for DNS whitelisting, whether deployment occurs
   universally or on an ad hoc basis.

The challenges of email whitelists and blacklists should be cited, since it
provides a rich base of experience for such an effort, at scale.


6.1.  Deploying DNS Whitelisting On An Ad Hoc Basis

   The seemingly most likely deployment scenario is where some

Most likely?  This is not already established practice?


   authoritative DNS server operators implement DNS whitelisting but
   many or most others do not do so.  What can make this scenario
   challenging from the standpoint of a DNS recursive resolver operator
   is determining which domains implement DNS whitelisting, particularly
   since a domain may not do so as they initially transition to IPv6,
   and may instead do so later.  Thus, a DNS recursive resolver operator



Livingood                Expires August 26, 2011               [Page 13]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   may initially believe that they can receive AAAA responses as a
   domain adopts IPv6, but then notice via end user reports that they no
   longer receive AAAA responses due to that domain adopting DNS
   whitelisting.  Of course, a domain's IPv6 transition may be
   effectively invisible to recursive server operators due to the effect
   of DNS whitelisting.

This suggests that every listing at the server needs a contact record for
periodic checks whether to renew the listing.



   In contrast to a universal deployment of DNS whitelisting
   Section 6.2, deployment on an ad hoc basis is likely to be
   significantly more challenging from an operational, monitoring, and

Oh?  Use in small scale is more challenging than use of manual exceptions list
at large scale?  That's a very unexpected view.


   troubleshooting standpoint.  In this scenario, a DNS recursive
   resolver operator will have no way to systematically determine
   whether DNS whitelisting is or is not implemented for a domain, since
   the absence of AAAA resource records may simply be indicative that
   the domain has not yet added IPv6 addressing for the domain, rather
   than that they have done so but have restricted query access via DNS

The premise is that, in large scale use, servers /will/ have a way to
systematically determine whether it is implemented?  What are the existing
examples of having such a capability for other Internet protocols and services?


   whitelisting.  As a result, discovering which domains implement DNS
   whitelisting, in order to differentiate them from those that do not,
   is likely to be challenging.

   One benefit of DNS whitelisting being deployed on an ad hoc basis is
   that only the domains that are interested in doing so would have to
   upgrade their authoritative DNS servers in order to implement the
   ACLs necessary to perform DNS whitelisting.

   In this potential deployment scenario, it is also possible that a
   given domain will implement DNS whitelisting temporarily.  A domain,
   particularly a highly-trafficked domain, may choose to do so in order
   to ease their transition to IPv6 through a selective deployment and
   minimize any perceived risk in such a transition.

6.2.  Deploying DNS Whitelisting Universally

   The least likely deployment scenario is one where DNS whitelisting is
   implemented on all authoritative DNS servers, across the entire
   Internet.  While this scenario seems less likely than ad hoc
   deployment due to some parties not sharing the concerns that have so
   far motivated the use of DNS whitelisting, it is nonetheless
   conceivable that it could be one of the ways in which DNS
   whitelisting is deployed.

Significantly, the partial-deployment model casts this mechanism as a transition
expedient -- as the document reasonably describes it -- whereas universal
deployment casts it as a fundamental change to the architecture.

Given that it would take decades to achieve relatively full deployment of this
'across the entire Internet', what is the benefit of discussing this highly
unlikely scenario?  Is it really "conceivable"?  I doubt it. If you think
otherwise, the paper needs to explore the deployment and adoption issues in much
more detail, because I don't see how it could work.


   In order for this deployment scenario to occur, it is likely that DNS
   whitelisting functionality would need to be built into all
   authoritative DNS server software, and that all operators of
   authoritative DNS servers would have to upgrade their software and
   enable this functionality.  It is likely that new Internet Draft
   documents would need to be developed which describe how to properly
   configure, deploy, and maintain DNS whitelisting.  As a result, it is



Livingood                Expires August 26, 2011               [Page 14]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   unlikely that DNS whitelisting would, at least in the next several
   years, become universally deployed.  Furthermore, these DNS
   whitelists are likely to vary on a domain-by-domain basis, depending
   upon a variety of factors.  Such factors may include the motivation
   of each domain owner, the location of the DNS recursive resolvers in
   relation to the source content, as well as various other parameters
   that may be transitory in nature, or unique to a specific end user
   host type.  It is probably unlikely that a single clearinghouse for
   managing whitelisting is possible; it will more likely be unique to
   the source content owners and/or domains which implement DNS
   whitelists.

   While this scenario may be unlikely, it may carry some benefits.
   First, parties performing troubleshooting would not have to determine
   whether or not DNS whitelisting was being used, as it always would be
   in use.  In addition, if universally deployed, it is possible that
   the criteria for being added to or removed from a DNS whitelist could
   be standardized across the entire Internet.  Nevertheless, even if
   uniform DNS whitelisting policies were not standardized, is also
   possible that a central registry of these policies could be developed
   and deployed in order to make it easier to discover them, a key part
   of achieving transparency regarding DNS whitelisting.

Is any of this paragraph realistic?  Obviously my asking means I don't it is.
These seem to be of theoretical rather than pragmatic interest.  ("If everyone
refuses to shoot, there will be no wars.")

It's true that this is an "implications" paper rather than a BCP, but still...



7.  Implications of DNS Whitelisting

   There are many potential implications of DNS whitelisting.  The key
   potential implications are detailed below.

7.1.  Architectural Implications

   DNS whitelisting could be perceived as modifying the end-to-end model
   and/or the general notion of the architecture that prevails on the

I'll suggest that perception is not a major issue about a technical topic like
this.  (It's not entirely irrelevant, of course, but I suspect it is quite minor.)

The major issue is whether it /actually/ modifies the end-to-end nature of the
DNS.  And I think it does, as well as modifying the "spontaneous
interoperability" expectation for most Internet mechanism, since it requires
prior registration.


7.2.  Public IPv6 Address Reachability Implications

   The predominant experience of end user hosts and servers on the IPv4-
   addressed Internet today is that when a new server with a public IPv4
   address is added to the DNS, that it is then globally accessible by

This sentence is not quite correct, in strict technical terms.  Since this is a
technical discussion, we need to be precise:  the host is reachable when the
routing tables make it reachable.  That's strictly a mapper of IP Address
handling, not name-to-address mapping.

What you mean is that its domain name is immediately useful for reaching it.


   IPv4-addressed hosts.  This is a generalization and in Section 5
   there are examples of common cases where this may not necessarily be
   the case.  For the purposes of this argument, that concept of
   accessibility can be considered "pervasive reachability".  It has so
   far been assumed that the same expectations of pervasive reachability
   would exist in the IPv6-addressed Internet.  However, if DNS
   whitelisting is deployed, this will not be the case since only end
   user hosts using DNS recursive resolvers which are included in the

again, you mean /name-based/ reachability.





Livingood                Expires August 26, 2011               [Page 16]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   ACL of a given domain using DNS whitelisting would be able to reach
   new servers in that given domain via IPv6 addresses.  The expectation
   of any end user host being able to connect to any server (essentially
   both hosts, just at either end of the network), defined here as
   "pervasive reachability", will change to "restricted reachability"
   with IPv6.

   Establishing DNS whitelisting as an accepted practice in the early
   phases of mass IPv6 deployment could well establish it as an integral
   part of how IPv6 DNS resource records are deployed globally.  As a
   result, it is then possible that DNS whitelisting could live on for
   decades on the Internet as a key foundational element of domain name
   management that we will all live with for a long time.

(that last sentence could benefit from some editing.)


   It is a critical to understand that the concept of reachability
   described above depends upon a knowledge or awareness of an address
   in the DNS.  Thus, in order to establish reachability to an end
   point, a host is dependent upon looking up an IP address in the DNS

If this section were started with a sentence like this, then there would not be
a problem with the other references' being confused with address-based routing
reachability.


   when a FQDN is used.  When DNS whitelisting is used, it is quite
   likely the case that an IPv6-enabled end user host could ping or
   connect to an example server host, even though the FQDN associated
   with that server host is restricted via a DNS whitelist.  Since most

First, I suspect that "example" doesn't add meaning to the sentence.  Second,
pinging and connecting might happen with or without the whitelist entry.  So I
do not understand what import there is in this sentence.


   Internet applications and hosts such as web servers depend upon the
   DNS, and as end users connect to FQDNs such as www.example.com and do
   not remember or wish to type in an IP address, the notion of
   reachability described here should be understood to include knowledge
   how to associate a name with a network address.

Again, this 'premise' statement should introduce the sub-section, not end it.



7.3.  Operational Implications

   This section explores some of the operational implications which may
   occur as a result of, are related to, or become necessary when
   engaging in the practice of DNS whitelisting.

7.3.1.  De-Whitelisting May Occur

The more general version of this issue is 'synchronization'.  Entries in the
whitelist need to be synchronized with host status and capabilities.


   It is possible for a DNS recursive resolver added to a whitelist to
   then be removed from the whitelist, also known as de-whitelisting.
   Since de-whitelisting can occur, through a decision by the
   authoritative server operator, the domain owner, or even due to a
   technical error, an operator of a DNS recursive resolver will have
   new operational and monitoring requirements and/or needs as noted in
   Section 7.3.3, Section 7.3.4, Section 7.3.6, and Section 7.5.

7.3.2.  Authoritative DNS Server Operational Implications

   Operators of authoritative servers may need to maintain an ACL a

a -> on a (?)


   server-wide basis affecting all domains, on a domain-by-domain basis,


Livingood                Expires August 26, 2011               [Page 17]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   as well as on a combination of the two.  As a result, operational

I'm not really understanding the first sentence.  One problem might be that its
discussing an implication of some configuration or usage options that have not
been previously specified, so that the reference here might be overly cryptic.

For example, I don't know what "affecting all domains" actually means.  It
almost sounds as if it could mean "everyone gets AAAA records" or "no one gets
AAAA records" yet I'm reaonably certain that is /not/ what is meant.


   practices and software capabilities may need to be developed in order
   to support such functionality.  In addition, processes may need to be
   put in place to protect against inadvertently adding or removing IP
   addresses, as well as systems and/or processes to respond to such
   incidents if and when they occur.  For example, a system may be
   needed to record DNS whitelisting requests, report on their status
   along a workflow, add IP addresses when whitelisting has been
   approved, remove IP addresses when they have been de-whitelisted, log
   the personnel involved and timing of changes, schedule changes to
   occur in the future, and to roll back any inadvertent changes.

Might be worth starting with a simple, broad summary statement, possibly along
the lines of:

   An AAAA DNS Whitelist serves as a critical infrastructure service; to be
useful it needs careful and extensive administration, monitoring and operation.
 Each new and essential mechanism creates substantial follow-on support costs.


   Operators may also need implement new forms of monitoring in order to
   apply change control, as noted briefly in Section 7.3.4.

7.3.3.  DNS Recursive Resolver Server Operational Implications

   Operators of DNS recursive resolvers, which may include ISPs,
   enterprises, universities, governments, individual end users, and
   many other parties, are likely to need to implement new forms of
   monitoring, as noted briefly in Section 7.3.4.  But more critically,
   such operators may need to add people, processes, and systems in
   order to manage large numbers of DNS whitelisting applications as
   part of their own IPv6 transition, for all domains that the end users
   of such servers are interested in now or in which they may be

I think the summary observation is simple and should be stated directly:  This
is a manual mechanism that becomes expensive in time and personnel effort as it
scales up.


   interested in the future.  As anticipation of interesting domains is
   likely infeasible, it is more likely that operators may either choose
   to only apply to be whitelisted for a domain based upon one or more
   end user requests, or that they will attempt to do so for all domains
   that they can ascertain to be engaging in DNS whitelisting.

"attempt to do so for all domain that they can ascertain to be engaging in DNS
whitelisting"  appears to be saying to do whitelisting for domains that do
whitelisting.  I don't understand.



   When operators apply for DNS whitelisting for all domains, that may

"apply for DNS whitelisting for all domains" -- again I'm not understanding what
this means.


7.3.5.  Implications of Operational Momentum

   It seems plausible that once DNS whitelisting is implemented it will
   be very difficult to deprecate such technical and operational
   practices.  This assumption is based in an understanding of human

in -> on


   nature, not to mention physics.  For example, as Sir Issac Newton
   noted, "Every object in a state of uniform motion tends to remain in
   that state of motion unless an external force is applied to it" [Laws

Code does not have momenum.  Neither do configurations or lists.  This really
isn't about physics.

It is entirely about group psychology, as you note, and the administrative
challenges in the logistics of large-scale operational changes (which probably
/does/ have something to with physics, but it seems a stretch to credit Newton.
How about Heisenberg?...)


   of Motion].  Thus, once DNS whitelisting is implemented it is quite
   likely that it would take considerable effort to deprecate the
   practice and remove it everywhere on the Internet - it will otherwise
   simply remain in place in perpetuity.  To better illustrate this
   point, one could consider one example (of many) that there are many
   email servers continuing to attempt to query or otherwise check anti-



Livingood                Expires August 26, 2011               [Page 19]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   spam DNS blocklists which have long ago ceased to exist.

7.3.6.  Troubleshooting Implications

   The implications of DNS whitelisted present many challenges, which
   have been detailed in Section 7.  These challenges may negatively

But this is /still/ section 7.  Can you be more specific?  Or perhaps say
"throughout this section".


   affect the end users' ability to troubleshoot, as well as that of DNS
   recursive resolver operators, ISPs, content providers, domain owners
   (where they may be different from the operator of the authoritative
   DNS server for their domain), and other third parties.  This may make
   the process of determining why a server is not reachable
   significantly more complex.

7.3.7.  Additional Implications If Deployed On An Ad Hoc Basis

   Additional implications, should this be deployed on an ad hoc basis,
   could include scalability problems relating to operational processes,

I'm pretty sure that scaling problems for this exist in all scenarios, not just
ad hoc usage.


   monitoring, and ACL updates.  In particular, it seems likely that as
   the number of domains that are using DNS whitelisting increases, as
   well as the number of IPv6-capable networks requesting to be
   whitelisted, that there is an increased likelihood of configuration
   and other operational errors, especially with respect to the ACLs
   themselves.

   It is unclear when and if it would be appropriate to change from
   whitelisting to blacklisting, and whether or how this could feasibly
   be coordinated across the Internet, which may be proposed or

Actually the question of coordination is quite clear and rather fundamental:

     No.

Anyone believing otherwise needs to cite a successful example, at Internet scale
and diversity, more recently than the 1983 switch to IP (which didn't go all
that well anyhow...)

Simple, unambiguous showstoppers should be stated in a simple and direct manner.
 When there is room for debate, softer language makes sense.  Again, if the
question of coordination really is subject to debate, then the basis needs to be
stated.  (Good luck!)


   implemented on an ad hoc basis when a majority of networks (or
   allocated IPv6 address blocks) have been whitelisted.  Finally, some
   parties implementing DNS whitelisting consider this to be a temporary
   measure.  As such, it is not clear how these parties will judge the
   network conditions to have changed sufficiently to justify disabling
   DNS whitelisting and/or what the process and timing will be in order
   to discontinue this practice.

   One further potential implication is that an end user with only an
   IPv4 address, using a DNS resolver which has not been whitelisted by
   any domains, would not be able to get any AAAA resource records.  In
   such a case, this could give that end user the incorrect impression
   that there is no IPv6-based content on the Internet since they are
   unable to discover any IPv6 addresses via the DNS.

7.4.  Homogeneity May Be Encouraged

   A broad trend which has existed on the Internet appears to be a move
   towards increasing levels of heterogeneity.  One manifestation of

increasing levels of heterogeneity -> more heterogeneity

(I think heterogeneity does not have 'levels'.)

Substantively:  say the nature of the heterogeneity within the initial claim.
For example, there is /less/ heterogeneity of ISPs, given industry
consolidation.  There is less heterogeneity of infrastructure equipment such as
routers.  Etc.


   this is in an increasing number, variety, and customization of end
   user hosts, including home network, operating systems, client



Livingood                Expires August 26, 2011               [Page 20]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


   software, home network devices, and personal computing devices.  This
   trend appears to have had a positive effect on the development and
   growth of the Internet.  A key facet of this that has evolved is the
   ability of the end user to connect any technically compliant device
   or use any technically compatible software to connect to the
   Internet.  Not only does this trend towards greater heterogeneity
   reduce the control which is exerted in the middle of the network,
   described in positive terms in [Tussle in Cyberspace], [Rethinking
   the Internet], and [RFC3724], but it can also help to enable greater
   and more rapid innovation at the edges.

   An unfortunate implication of the adoption of DNS whitelisting may be
   the encouragement of a reversal of this trend, which would be a move

the encouragement of -> to encourage


8.1.  Implement DNS Whitelisting Universally

   One obvious solution is to implement DNS whitelisted universally, and
   to do so using some sort of centralized registry of DNS whitelisting
   policies, contracts, processes, or other information.  This potential
   solution seems unlikely at the current time.

I'm pretty sure that the only thing that is obvious about a premise of universal
adoption is that it's not practical.  Seriously.

At the least, this section needs to be less cavalier about putting this
alternative forward as a "solution", especially given the rather serious
drawbacks/problems with it.


8.2.  Implement DNS Whitelisting On An Ad Hoc Basis

   If DNS whitelisting is to be adopted, it is likely to be adopted on

"is to be"?  I thought it already had a significant installed base.


   this ad hoc, or domain-by-domain basis.  Therefore, only those
   domains interested in DNS whitelisting would need to adopt the
   practice, though as noted herein discovering that they a given domain
   has done so may be problematic.  Also in this scenario, ad hoc use by
   a particular domain may be a temporary measure that has been adopted
   to ease the transition of the domain to IPv6 over some short-term
   timeframe.

8.3.  Do Not Implement DNS Whitelisting

   As an alternative to adopting DNS whitelisting, the Internet
   community generally can choose to take no action whatsoever,
   perpetuating the current predominant authoritative DNS operational
   model on the Internet, and leave it up to end users with IPv6-related
   impairments to discover and fix those impairments.


That is, place the burden of fixing a problem on those creating it?






Livingood                Expires August 26, 2011               [Page 23]

Internet-Draft   IPv6 AAAA DNS Whitelisting Implications   February 2011


8.3.1.  Solving Current End User IPv6 Impairments

   A further extension of not implementing DNS whitelisting, is to also
   endeavor to actually fix the underlying technical problems that have
   prompted the consideration of DNS whitelisting in the first place, as
   an alternative to trying to apply temporary workarounds to avoid the
   symptoms of underlying end user IPv6 impairments.  A first step is
   obviously to identify which users have such impairments, which would
   appear to be possible, and then to communicate this information to
   end users.  Such end user communication is likely to be most helpful
   if the end user is not only alerted to a potential problem but is
   given careful and detailed advice on how to resolve this on their
   own, or where they can seek help in doing so.  Section 11 may also be
   relevant in this case.

   One challenge with this option is the potential difficulty of
   motivating members of the Internet community to work collectively
   towards this goal, sharing the labor, time, and costs related to such
   an effort.  Of course, since just such a community effort is now
   underway for IPv6, it is possible that this would call for only a
   moderate amount of additional work.

This 'challenge' is at the core of /all/ adoption efforts for Internet protocols
and services that entail distributed adoption.


   Despite any potential challenges, many in the Internet community are
   already working towards this goal and/or have expressed a general
   preference for this approach.

If this is not already an organized effort with a website, sponsoring
consortium, or the like, it should be.  If it is, then cite it in this doc!


8.3.2.  Gain Experience Using IPv6 Transition Names

   Another alternative is for domains to gain experience using an FQDN
   which has become common for domains beginning the transition to IPv6;
   ipv6.example.com and www.ipv6.example.com.  This can be a way for a
   domain to gain IPv6 experience and increase IPv6 use on a relatively
   controlled basis, and to inform any plans for DNS whitelisting with
   experience.

I do not understand what this means.

What is it for?  What are the results?  How are theyused?


9.  Is DNS Whitelisting a Recommended Practice?

   Opinions in the Internet community concerning whether or not DNS
   whitelisting is a recommended practice are understandably quite
   varied.  However, there is clear consensus that DNS whitelisting is
   at best a useful temporary measure which a domain may choose to

If that is a clear consensus, then it makes even less sense to promote the idea
of universal adoption, given the timescale needed to achieve it.


10.  Security Considerations

   There are no particular security considerations if DNS whitelisting
   is not adopted, as this is how the public Internet works today with A
   resource records.

Or rather, failure to adopt a mechanism like this or repair the underlying
problem, for those sites experiencing that problem, will result in a denial of
service, albeit not an intentional one.  Still, that's a pretty basic security
issue.


d/
--

  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf


--

  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf



[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]