<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE rfc [
<!ENTITY nbsp    "&#160;">
<!ENTITY zwsp    "&#8203;">
<!ENTITY nbhy    "&#8209;">
<!ENTITY wj      "&#8288;">
]>
<rfc xmlns:xi="http://www.w3.org/2001/XInclude"
     category="info"
     docName="draft-sun-single-stack-100-50-00"
     ipr="trust200902"
     obsoletes=""
     updates=""
     submissionType="independent"
     version="3">

  <front>
    <title abbrev="Single-Stack 100/50 Principle">The Single-Stack 100/50 Principle: Formal Definitions for IPv4 Retirement in Dual-Stack Networks</title>

    <seriesInfo name="Internet-Draft" value="draft-sun-single-stack-100-50-00"/>

    <author fullname="Charles C. Sun" initials="C." surname="Sun">
      <organization>Alliance for Universal Computing</organization>
      <address>
        <postal>
          <city>Potomac</city>
          <region>Maryland</region>
          <country>United States of America</country>
        </postal>
        <email>charles.sun@aucglobal.com</email>
      </address>
    </author>

    <date year="2026" month="September" day="2"/>

    <area>General</area>
    <workgroup>Network Working Group</workgroup>

    <keyword>IPv4 retirement</keyword>
    <keyword>IPv6</keyword>
    <keyword>dual-stack</keyword>
    <keyword>attack surface</keyword>

    <abstract>
      <t>This document defines two formally provable consequences of
      retiring the IPv4 protocol stack in a dual-stack (IPv4 and IPv6)
      network environment: (1) complete elimination of attacks
      attributable to the IPv4 protocol, and (2) a minimum 50% reduction
      in the count of concurrently exposed network-layer attack
      surfaces. Both results are established as definitional and
      structural derivations, independent of empirical attack-volume
      measurement. This document proposes standard terminology for
      discussing these effects and states explicitly the boundary
      between what is proven here and what remains an open empirical
      question.</t>
    </abstract>

  </front>

  <middle>

    <section anchor="introduction" numbered="true" toc="default">
      <name>Introduction</name>
      <t>Dual-stack operation of IPv4 and IPv6 is widely deployed as a
      transition strategy toward IPv6-only networking. This document
      states, in formal terms, two consequences that follow directly
      from retiring the IPv4 stack in such an environment. Both
      consequences are established by definition and by direct
      enumeration; neither depends on measured attack-volume data. A
      separate empirical question -- the relative share of attack
      volume historically carried by each stack -- is explicitly out
      of scope for the claims made in <xref target="theorem-i"/> and
      <xref target="theorem-ii"/>, and is discussed in
      <xref target="scope"/>.</t>
    </section>

    <section anchor="terminology" numbered="true" toc="default">
      <name>Terminology</name>
      <t>This document defines no conformance requirements and uses
      no RFC&nbsp;2119 key words; the terms below are used in their
      ordinary English sense.</t>

      <dl newline="false" spacing="normal">
        <dt>Dual-Stack Mode of Operations:</dt>
        <dd>a network or host configuration in which both the IPv4 and
        IPv6 protocol stacks are concurrently active and reachable.</dd>

        <dt>IPv4-Attributable Attack:</dt>
        <dd>an attack vector whose execution requires the IPv4
        protocol, including but not limited to IPv4 address-space
        scanning, IPv4 NAT-traversal exploitation, ARP spoofing, IPv4
        header/fragmentation manipulation, and IPv4-specific route
        hijacking.</dd>

        <dt>Network-Layer Attack Surface:</dt>
        <dd>a complete, independently addressable protocol stack
        presenting its own addressing scheme, discovery/resolution
        mechanism, header format, and protocol-specific vulnerability
        set.</dd>

        <dt>Single-Stack Mode of Operations:</dt>
        <dd>a network or host configuration in which only one
        network-layer protocol stack (IPv4 or IPv6) is active and
        reachable.</dd>
      </dl>
    </section>

    <section anchor="theorem-i" numbered="true" toc="default">
      <name>Theorem I: Complete Elimination of IPv4-Attributable Attacks</name>
      <t>By definition (<xref target="terminology"/>), an
      IPv4-Attributable Attack requires the IPv4 stack to be present
      and reachable. If the IPv4 stack does not exist on a network, no
      packet using IPv4 addressing or IPv4 protocol mechanics can
      reach a host on that network.</t>

      <t><strong>Statement:</strong> Retiring the IPv4 stack yields
      complete (100%) elimination of IPv4-Attributable Attacks against
      that network.</t>

      <t><strong>Justification:</strong> This follows directly from
      the definition of "IPv4-Attributable Attack" in
      <xref target="terminology"/> and requires no supporting
      measurement.</t>
    </section>

    <section anchor="theorem-ii" numbered="true" toc="default">
      <name>Theorem II: Minimum 50% Reduction in Attack Surface Count</name>
      <t>In Dual-Stack Mode of Operations, a host or network
      concurrently exposes two independent, complete Network-Layer
      Attack Surfaces: the IPv4 surface and the IPv6 surface. Let S
      denote the count of concurrently exposed Network-Layer Attack
      Surfaces.</t>

      <artwork name="" type="" align="left" alt="">
S(dual-stack)    = 2   (IPv4 surface + IPv6 surface)
S(IPv4 retired)  = 1   (IPv6 surface only)
      </artwork>

      <t><strong>Statement:</strong> Retiring the IPv4 stack reduces S
      by (2&nbsp;-&nbsp;1)&nbsp;/&nbsp;2&nbsp;=&nbsp;50%, taken as a
      minimum.</t>

      <t><strong>Justification:</strong> The reduction is stated as a
      minimum because the surviving IPv6 surface continues to face
      attacks at up to 100% of its own scope; that 100% is now the
      entirety of a single-stack environment, not one term of a
      two-stack total. The reduction in S holds regardless of the
      relative attack volume previously carried by each stack, because
      S counts exposed surfaces, not traffic or incidents.</t>
    </section>

    <section anchor="scope" numbered="true" toc="default">
      <name>Scope and Non-Goals</name>
      <t>Theorems I and II in <xref target="theorem-i"/> and
      <xref target="theorem-ii"/> are definitional and structural.
      They do not assert, and this document takes no position on, the
      separate empirical question of what percentage of overall global
      attack volume or incident count is currently attributable to
      IPv4 versus IPv6. Any such claim would require direct measurement
      of comparative attack volume across both stacks, which is not
      undertaken in this document.</t>
    </section>

    <section anchor="security-considerations" numbered="true" toc="default">
      <name>Security Considerations</name>
      <t>This document defines terminology and formal properties
      related to protocol retirement; it specifies no new protocol
      mechanism, message format, or implementable behavior, and
      introduces no new attack surface of its own. Operators retiring
      IPv4 in favor of IPv6-only operation should independently
      evaluate IPv6-specific security considerations (e.g., Neighbor
      Discovery security, extension header handling, and
      address-scanning resistance) which are outside the scope of this
      document.</t>
    </section>

    <section anchor="iana-considerations" numbered="true" toc="default">
      <name>IANA Considerations</name>
      <t>This document has no IANA actions.</t>
    </section>

  </middle>

  <back>
  </back>

</rfc>
