Network Working Group O. Troan, Ed.
Internet-Draft cisco
Intended status: Standards Track October 31, 2011
Expires: May 3, 2012
Mapping of Address and Port (MAP)
draft-mdt-softwire-mapping-address-and-port-01
Abstract
This document describes a generic mechanism for mapping between an
IPv4 prefix, address or parts thereof, and transport layer ports and
an IPv6 prefix or address.
Status of this Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on May 3, 2012.
Copyright Notice
Copyright (c) 2011 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Troan Expires May 3, 2012 [Page 1]
Internet-Draft MAP October 2011
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 6
3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 7
4. Mapping Rules . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1. Port mapping algorithm . . . . . . . . . . . . . . . . . . 10
4.1.1. Bit Representation of the Algorithm . . . . . . . . . 11
4.1.2. GMA examples . . . . . . . . . . . . . . . . . . . . . 11
4.1.3. GMA Provisioning Considerations . . . . . . . . . . . 12
4.1.4. Features of the Algorithm . . . . . . . . . . . . . . 12
4.2. Basic mapping rule (BMR) . . . . . . . . . . . . . . . . . 13
4.3. Forwarding mapping rule (FMR) . . . . . . . . . . . . . . 15
4.4. Default mapping rule (DMR) . . . . . . . . . . . . . . . . 16
5. Use of the IPv6 Interface identifier . . . . . . . . . . . . . 18
6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20
7. Security Considerations . . . . . . . . . . . . . . . . . . . 21
8. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 22
9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 23
10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 24
10.1. Normative References . . . . . . . . . . . . . . . . . . . 24
10.2. Informative References . . . . . . . . . . . . . . . . . . 24
Appendix A. Open issues / New features . . . . . . . . . . . . . 28
A.1. Max PSID . . . . . . . . . . . . . . . . . . . . . . . . . 28
A.2. Interface identifier - V octet and Checksum neutrality . . 28
A.3. Optional BR per Rule within a domain . . . . . . . . . . . 29
Appendix B. Requirements . . . . . . . . . . . . . . . . . . . . 30
Appendix C. Deployment considerations . . . . . . . . . . . . . . 32
C.1. Flexible Assigment of Port Sets . . . . . . . . . . . . . 32
C.2. Traffic Classification . . . . . . . . . . . . . . . . . . 32
C.3. Prefix Delegation Deployment . . . . . . . . . . . . . . . 32
C.4. Coexisting Deployment . . . . . . . . . . . . . . . . . . 32
C.5. Friendly to Network Provisioning . . . . . . . . . . . . . 33
C.6. Enable privacy addresses . . . . . . . . . . . . . . . . . 33
C.7. Facilitating 4v6 Service . . . . . . . . . . . . . . . . . 33
C.8. Independency with IPv6 Routing Planning . . . . . . . . . 33
C.9. Optimized Routing Path . . . . . . . . . . . . . . . . . . 33
Appendix D. Guidelines for Operators . . . . . . . . . . . . . . 34
D.1. Additional terms . . . . . . . . . . . . . . . . . . . . . 34
D.2. Understanding address formats: their difference and
relevance . . . . . . . . . . . . . . . . . . . . . . . . 34
D.3. Residual deployment with MAP . . . . . . . . . . . . . . . 38
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 42
Troan Expires May 3, 2012 [Page 2]
Internet-Draft MAP October 2011
1. Introduction
The mechanism of mapping IPv4 addresses in IPv6 address has been
described in numerous mechanisms dating back to [RFC1933] from 1996.
The Automatic tunneling mechanism described in RFC1933, assigned a
globally unique IPv6 address to a host by combining the hosts IPv4
address with a well known IPv6 prefix. Given an IPv6 packet with an
destination address with an embedded IPv4 address, a node could
automatically tunnel this packet by extracting the IPv4 tunnel end-
point address from the IPv6 destination address.
There are numerous variations of this idea, described in 6over4
[RFC2529], ISATAP [RFC5214] and 6rd [RFC5969]. The differences are
the use of well known IPv6 prefixes, or Service Provider assigned
IPv6 prefixes, and the exact position of the IPv4 bits embedded in
the IPv6 address. Teredo [RFC4380] added a twist to this to achieve
NAT traversal by also encoding transport layer ports into the IPv6
address. 6rd to achieve more efficient encoding, allowed for only an
IPv4 address suffix to be embedded, with the IPv4 prefix being
deducted from other provisioning mechanisms.
NAT-PT [RFC2766](deprecated) combined with a DNS ALG used address
mapping to put NAT state, namely the IPv6 to IPv4 binding encoded in
an IPv6 address. This characteristic has been inherited by NAT64
[RFC6146] and DNS64 [RFC6147] which rely on an address format defined
in [RFC6052]. [RFC6052] specifies the algorithmic translation of an
IPv6 address to IPv4 address suffix to be embedded, with the deducted
from other provisioning mechanisms. DNS ALG used address IPv4
binding encoded in it a corresponding IPv4 address, and vice versa.
In particular, [RFC6052] specifies the address format to build IPv4-
converted and IPv4-translatable IPv6 addresses. RFC6052 discusses
the transport of the port set information in an IPv4-embedded IPv6
address but the conclusion was the following (excerpt from
[RFC6052]):
"There have been proposals to complement stateless translation with a
port range feature. Instead of mapping an IPv4 address to exactly
one IPv6 prefix, the options would allow several IPv6 nodes to share
an IPv4 address, with each node managing a different set of ports.
If a port set extension is needed, could be defined later, using bits
currently reserved as null in the suffix."
The commonalities of all these mechanisms are:
o Provisions an IPv6 address for a host or an IPv6 prefix for a site
o Algorithmic or implicit address resolution for tunneling or
encapsulation. Given an IPv6 destination address, an IPv4 tunnel
Troan Expires May 3, 2012 [Page 3]
Internet-Draft MAP October 2011
endpoint address can be calculated. Likewise for translation, an
IPv4 address can be calculated from an IPv6 destination address
and vice versa.
o Embedding of an IPv4 address or part thereof and optionally
transport layer ports into an IPv6 address.
In the later phases of IPv4 to IPv6 migration, IPv6 only networks
will be common, while there will still be a need for residual IPv4
deployment. This document describes a more generic mapping of IPv4
to IPv6 that can be used both for encapsulation (IPv4 over IPv6) and
for translation between the two protocols.
Just as the IPv6 over IPv4 mechanisms refereed to above, the residual
IPv4 over IPv6 mechanisms must be capable of:
o Provisioning an IPv4 prefix, an IPv4 address or a shared IPv4
address.
o Algorithmically map between an IPv4 prefix, IPv4 address or a
shared IPv4 address and an IPv6 address.
The unified mapping scheme described here supports translation mode,
encapsulation mode, in both mesh and hub and spoke topologies.
This document describes delivery of IPv4 unicast service across an
IPv6 infrastructure. IPv4 multicast is not considered further in
this document.
Other work that has motivated the work on a unified mapping mechanism
for translation and encapsulation are:
[I-D.sun-softwire-stateless-4over6]
[I-D.murakami-softwire-4v6-translation]
[I-D.despres-softwire-4rd-addmapping]
[I-D.chen-softwire-4v6-add-format] [I-D.bcx-address-fmt-extension]
[I-D.mrugalski-dhc-dhcpv6-4rd]
[I-D.boucadair-dhcpv6-shared-address-option]
[I-D.despres-softwire-sam] [I-D.chen-softwire-4v6-pd]
[I-D.boucadair-softwire-stateless-requirements]
[I-D.dec-stateless-4v6] [I-D.boucadair-behave-ipv6-portrange]
[I-D.bsd-softwire-stateless-port-index-analysis]
[I-D.despres-softwire-stateless-analysis-tool]
[I-D.xli-behave-divi-pd] [I-D.murakami-softwire-4rd].
In particular the architecture of a shared IPv4 address by
distributing the port space is described in [RFC6346]. The
corresponding stateful solution DS-lite is described in [RFC6333]
Troan Expires May 3, 2012 [Page 4]
Internet-Draft MAP October 2011
Outstanding issues, Requirements and deployment considerations are
temporarily kept in Appendix A to D. The appendixes are in no way to
be considered normative.
Troan Expires May 3, 2012 [Page 5]
Internet-Draft MAP October 2011
2. Conventions
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119].
Troan Expires May 3, 2012 [Page 6]
Internet-Draft MAP October 2011
3. Terminology
MAP domain: A set of MAP CEs and BRs connected to the same
virtual link. A service provider may deploy a
single MAP domain, or may utilize multiple MAP
domains.
MAP Rule A set of parameters describing the mapping
between an IPv4 prefix, IPv4 address or shared
IPv4 address and an IPv6 prefix or address.
Each MAP node in the domain has the same set of
rules.
MAP Border Relay (BR): A MAP enabled router managed by the service
provider at the edge of a MAP domain. A Border
Relay router has at least an IPv6-enabled
interface and an IPv4 interface connected to
the native IPv4 network. A MAP BR may also be
referred to simply as a "BR" within the context
of MAP.
MAP Customer Edge (CE): A device functioning as a Customer Edge
router in a MAP deployment. In a residential
broadband deployment, this type of device is
sometimes referred to as a "Residential
Gateway" (RG) or "Customer Premises Equipment"
(CPE). A typical MAP CE adopting MAP rules
will serve a residential site with one WAN side
interface, one or more LAN side interfaces. A
MAP CE may also be referred to simply as a "CE"
within the context of MAP.
Shared IPv4 address: An IPv4 address that is shared among multiple
CEs. Each node has a separate part of the
transport layer port space; denoted as a port
set. Only ports that belong to the assigned
port set can be used for communication.
End-user IPv6 prefix: The IPv6 prefix assigned to an End-user CE by
other means than MAP itself.
MAP IPv6 address: The IPv6 address used to reach the MAP function
of a CE from other CE's and from BR's.
Port-set ID (PSID): Algorithmically identifies a set of ports
exclusively assigned to the CE.
Troan Expires May 3, 2012 [Page 7]
Internet-Draft MAP October 2011
Rule IPv6 prefix: An IPv6 prefix assigned by a Service Provider
for a mapping rule.
Rule IPv4 prefix: An IPv4 prefix assigned by a Service Provider
for a mapping rule.
IPv4 Embedded Address (EA) bits: The IPv4 EA-bits in the IPv6
address identify an IPv4 prefix/address (or
part thereof) or a shared IPv4 address (or part
thereof and a port set identifier.
Troan Expires May 3, 2012 [Page 8]
Internet-Draft MAP October 2011
4. Mapping Rules
A MAP node is provisioned with one or more mapping rules.
Mapping rules are used differently depending on their function.
Every MAP node must be provisioned with a Basic mapping rule. This
is used by the node to map from an End-user IPv6 prefix to an IPv4
prefix, address or shared IPv4 address. This same basic rule can
also be used for forwarding, where an IPv4 destination address and
optionally a destination port is mapped into an IPv6 address or
prefix. Additional mapping rules can be specified to allow for e.g.
multiple different IPv4 subnets to exist within the domain.
Additional mapping rules are recognized by having a Rule IPv6 prefix
different from the base End-user IPv6 prefix.
Traffic outside of the domain (IPv4 address not matching (using
longest matching prefix) any Rule IPv4 prefix in the Rules database)
will be forward using the Default Rule. The Default Rule maps
outside destinations to the BR's IPv6 address.
There are three types of mapping rules:
1. Basic Mapping Rule - used for IPv4 prefix, address or port set
assignment. There can only be one Basic Mapping Rule per End-
user IPv6 prefix.
* Rule IPv6 prefix (including prefix length)
* Rule IPv4 prefix (including prefix length)
* Rule EA-bits length (in bits)
* Rule Port Parameters (optional)
2. Forwarding Mapping Rule - used for forwarding. The Basic Mapping
Rule is also a Forwarding Mapping Rule. Each Forwarding Mapping
Rule will result in a route in a conceptual RIB for the Rule IPv4
prefix.
* Rule IPv6 prefix (including prefix length)
* Rule IPv4 prefix (including prefix length)
* Rule EA-bits length (in bits)
* Rule Port Parameters (optional)
Troan Expires May 3, 2012 [Page 9]
Internet-Draft MAP October 2011
3. Default Mapping Rule - used for destinations outside the MAP
domain. A 0.0.0.0/0 route is installed in the RIB for this rule.
* Rule IPv6 prefix (including prefix length)
* Rule BR IPv4 address
A MAP node finds its Basic Mapping Rule by doing a longest match
between the End-user IPv6 prefix and the Rule IPv6 prefix in the
Mapping Rule database. The rule is then used for IPv4 prefix,
address or shared address assignment.
Routes in the conceptual RIB are installed for all the Forwarding
Mapping Rules and an IPv4 default route for the Default Mapping Rule.
In the hub and spoke mode, all traffic should be forwarded using the
Default Mapping Rule.
4.1. Port mapping algorithm
Several port mapping algorithms have been proposed with their own set
of advantages and disadvantages. Since different PSID MUST have non-
overlapping port sets, the two extreme cases are: (1) the port number
is not contiguous for each PSID, but uniformly distributed across the
whole port range (0-65535); (2) the port number is contiguous in a
single range for each PSID. The port mapping algorithm proposed here
is called generalized modulus algorithm (GMA) and supports both these
cases.
For a given sharing ratio (R) and the maximum number of contiguous
ports (M), the GMA algorithm is defined as:
1. The port number (P) of a given PSID (K) is composed of:
P = R * M * j + M * K + i
Where:
* PSID: K = 0 to R - 1
* Port range index: j = (1024 / M) / R to ((65536 / M) / R) - 1,
if the well-known port numbers (0 - 1024) are excluded.
* Contiguous Port index: i = 0 to M - 1
2. The PSID (K) of a given port number (P) is determined by:
K = (floor(P/M)) % R
Troan Expires May 3, 2012 [Page 10]
Internet-Draft MAP October 2011
Where:
* % is the modulus operator
* floor(arg) is a function that returns the largest integer not
greater than arg
4.1.1. Bit Representation of the Algorithm
Given a sharing ratio (R=2^k), the maximum number of contiguous ports
(M=2^m), for any PSID (K) and available ports (P) can be represented
as:
0 8 15
+---------------+----------+------+-------------------+
| P |
----------------+-----------------+-------------------+
| A (j) | PSID (K) | M (i) |
+---------------+----------+------+-------------------+
|<----a bits--->|<-----k bits---->|<------m bits----->|
|k-c |<--c bits-->|<------m bits----->|
Figure 1: Bit representation
Where j and i are the same indexes defined in the port mapping
algorithm.
For any port number, the PSID can be obtained by bit mask operation.
Note that in above figure there is a PSID prefix length (c). Based
on this definition, PSID can also be represented in "CIDR style" and
more ports can be assigned to a single CE when PSID prefix length (c
< k).
When m = 0, GMA becomes a modulo operation. When a = 0, GMA becomes
division operation. The port mapping algorithm in
[I-D.despres-softwire-4rd-addmapping] can be represented by the
algorithm usng a=4 and each PSID may have different prefix length c).
4.1.2. GMA examples
Troan Expires May 3, 2012 [Page 11]
Internet-Draft MAP October 2011
For example, for R=128, M=4,
Port set-1 Port set-2
PSID=0 | 1024, 1025, 1026, 1027, | 1536, 1537, 1538, 1539, | 2048
PSID=1 | 1028, 1029, 1030, 1031, | 1540, 1541, 1542, 1543, | ....
PSID=2 | 1032, 1033, 1034, 1035, | 1544, 1545, 1546, 1547, | ....
PSID=3 | 1036, 1037, 1038, 1039, | 1548, 1549, 1550, 1551, | ....
...
PSID=127 | 1532, 1533, 1534, 1535, | 2044, 2045, 2046, 2047, | ....
Figure 2: Example
4.1.3. GMA Provisioning Considerations
The sharing ratio (R), the PSID (K) and the PSID length are derived
from existing information.
The number of offset bits (A) and excluded ports are optionally
provisioned via the "Rule Port Mapping Parameters" in the Basic
Mapping Rule.
The defaults are:
o Excluded ports : 0-1023
o Offset bits (A) : 6
The defaults of Offset bits (A), which determines excluded ports,
remains to be chosen. At least if MAP and native-IPv6 prefixes are
the same, two values are considered: 6 and 4. With offset=6, there
are 1024 excluded ports, but the maximum sharing ratio is less than
the requirement of R-4 (1024). With offset=4, compliance with R-4 is
ensured, but there are 4096 excluded ports, which reduces by 4.8% the
number of non-well-known ports that can be unused 4096-1024)/
(65536-1024). Comparative merits of R-4 compliance and full
optimization of port-set sizes remain to be evaluated. If MAP and
native-IPv6 prefixes are different, having a different default, e.g.
offset=0 has also been proposed.
4.1.4. Features of the Algorithm
The GMA algorithm has the following features:
1. There is no waste of the port numbers, except the well-known
ports.
Troan Expires May 3, 2012 [Page 12]
Internet-Draft MAP October 2011
2. The algorithm is flexible, the control parameters are sharing
ratio (R), the continue port range (M) and PSID prefix length
(c).
3. The algorithm is simple to perform effectively.
4. It allows Service Providers to define their own address sharing
ratio, the theoretical value is from 1:1 to 1:65536 and a more
practical value is from 1:1 to 1:4096.
5. It supports deployments using differentiated port ranges.
6. It could support differentiated port ranges within a single
shared IPv4 address, depending on the IPv6 format chosen (see
Appendix A).
7. It support excluding the well known ports 0-1023.
8. It supports assigning well known ports to a CE.
9. It supports legacy RTP/RTCP compatibility.
4.2. Basic mapping rule (BMR)
| n bits | o bits | m bits | 128-n-o-m bits |
+--------------------+-----------+---------+------------+----------+
| Domain IPv6 prefix | EA bits |subnet ID| interface ID |
+--------------------+-----------+---------+-----------------------+
|<--- End-user IPv6 prefix --->|
Figure 3: IPv6 address format
The Embedded Address bits (EA bits) are unique per end user within a
Domain IPv6 prefix. The Domain IPv6 prefix is the part of the End-
user IPv6 prefix that is common among all CEs using the same Basic
Mapping Rule within the MAP domain. There MUST be a Basic Mapping
Rule with a Rule IPv6 prefix equal to the Domain IPv6 prefix. The EA
bits encode the CE specific IPv4 address and port information. The
EA bits can contain a full or part of an IPv4 prefix or address, and
in the shared IPv4 address case contains a Port Set Identifier
(PSID).
Troan Expires May 3, 2012 [Page 13]
Internet-Draft MAP October 2011
Shared IPv4 address:
| r bits | p bits | | q bits |
+-------------+---------------------+ +------------+
| Domain IPv4 | IPv4 Address suffix | |Port Set ID |
+-------------+---------------------+ +------------+
| 32 bits |
Figure 4
Complete IPv4 address:
| r bits | p bits |
+-------------+---------------------+
| Domain IPv4 | IPv4 Address suffix |
+-------------+---------------------+
| 32 bits |
Figure 5
IPv4 prefix:
| r bits | p bits |
+-------------+---------------------+
| Domain IPv4 | IPv4 Address suffix |
+-------------+---------------------+
| < 32 bits |
Figure 6
If only a part of the IPv4 address/prefix is encoded in the EA bits,
the Domain IPv4 prefix is provisioned to the CE by other means (e.g.
a DHCPv6 option). To create a complete IPv4 address (or prefix), the
IPv4 address suffix from the EA bits, are concatenated with the
Domain IPv4 prefix (r bits).
The offset of the EA bits field in the IPv6 address is equal to the
BMR Rule IPv6 prefix length. The length of the EA bits field (o) is
given in the Rule EA-bits length parameter.
If o + r < 32, then an IPv4 prefix is assigned. The IPv4 prefix
length is equal to r bits + Rule EA-bits length.
Troan Expires May 3, 2012 [Page 14]
Internet-Draft MAP October 2011
If o + r is equal to 32, then a full IPv4 address is to be assigned.
The address is created by concatenating the Domain IPv4 prefix and
the EA-bits.
If o + r is > 32, then a shared IPv4 address is to be assigned. The
number of IPv4 address bits (p) in the EA bits is given by 32 - r
bits. The PSID bits are used to create a port set. The length of
the PSID bit field within EA bits is: o - p.
| Port range (16 bits) |
+---------------+----------+------+-------------------+
| P |
----------------+-----------------+-------------------+
| A (j) | PSID (K) | M (i) |
+---------------+----------+------+-------------------+
|<----a bits--->|<-----k bits---->|<------m bits----->|
|<---c bits--->|<-----(k+m-c) bits--->|
Figure 7
Example:
Given:
End-user IPv6 prefix: 2001:db8:0012:34::/56
Domain IPv6 prefix: 2001:db8:00::/40
IPv4 prefix: 192.0.2.0/24
Basic Mapping Rule: {2001:db8:00::/40, 192.0.2.0/24, 256, 6}
We get IPv4 address and port set:
EA bits offset: 40
IPv4 suffix bits (p): 32 - 24 = 8
IPv4 address: 192.0.2.18
PSID start: 40 + p = 40 + 8 = 48
PSID length: o - p = log2(256) - 8 = 8.
PSID: 0x34.
4.3. Forwarding mapping rule (FMR)
On adding a FMR rule an IPv4 route is installed the RIB (conceptual)
for the Rule IPv4 prefix.
On forwarding an IPv4 packet a lookup is done in the RIB and the
correct FMR is used.
Troan Expires May 3, 2012 [Page 15]
Internet-Draft MAP October 2011
| 32 bits | | 16 bits |
+--------------------------+ +-------------------+
| IPv4 destination address | | IPv4 dest port |
+--------------------------+ +-------------------+
: : ___/ :
| p bits | / q bits :
+----------+ +------------+
|IPv4 sufx| |Port Set ID |
+----------+ +------------+
\ / ____/ ________/
\ : __/ _____/
\ : / /
| n bits | o bits | m bits | 128-n-o-m bits |
+--------------------+-----------+---------+------------+----------+
| Domain IPv6 prefix | EA bits |subnet ID| interface ID |
+--------------------+-----------+---------+-----------------------+
|<--- End-user IPv6 prefix --->|
Figure 8
The subnet ID for MAP is defined to be ~0. I.e. the last subnet in
an End-user IPv6 prefix allocation is used for MAP. A MAP node MUST
reserve the topmost IPv6 prefix in a End-user IPv6 prefix for the
purpose of MAP. This prefix MUST NOT be used for native IPv6
traffic.
Example:
Given:
IPv4 destination address: 192.0.2.18
IPv4 destination port: 1232
Forwarding Mapping Rule: {2001:db8:00::/40, 192.0.2.0/24,
Sharing ratio: 256, PSID offset: 6}
We get IPv6 address:
IPv4 suffix bits (p): 32 - 24 = 8 (18)
PSID length: 8 (sharing ratio)
PSID: 0x34 (1232)
EA bits: 0x1234
IPv6 address: 2001:db8:0012:34FF:<interface-identifier>
4.4. Default mapping rule (DMR)
The Default Mapping rule is used to reach IPv4 destinations outside
of the MAP domain. Traffic using this rule will be sent from a CE to
a BR.
Troan Expires May 3, 2012 [Page 16]
Internet-Draft MAP October 2011
The Rule IPv4 prefix in the DMR is: 0.0.0.0/0. The Rule IPv6 prefix
is the IPv6 address or prefix of the BR. Which is used is dependent
on the mode used. For example translation requires that the IPv4
destination address is encoded in the BR IPv6 address, so only a
prefix is used in the DMR to allow for a generated interface
identifier. For the encapsulation mode the Rule IPv6 prefix can be
the full IPv6 address of the BR.
An example of a DMR is:
Default Mapping Rule: {2001:db8:0001:0000:<interface-id>:/128,
0.0.0.0/0, BR IPv4 address: 192.0.2.1, }
In most implementations of a RIB, the next-hop address must be of the
same address family as the prefix. To satisfy this requirement a BR
IPv4 address is included in the rule. Giving a default route in the
RIB:
0.0.0.0 -> 192.0.2.1, MAP-Interface0
Troan Expires May 3, 2012 [Page 17]
Internet-Draft MAP October 2011
5. Use of the IPv6 Interface identifier
In an encapsulation solution, an IPv4 address and port is mapped to
an IPv6 address. This is the address of the tunnel end point of the
receiving MAP CE. For traffic outside the MAP domain, the IPv6
tunnel end point address is the IPv6 address of the BR. As long as
the interface-id is well known or provisioned and the same for all
MAP nodes, it can be any interface identifier. E.g. ::1.
When translating, the destination IPv4 address is translated into a
corresponding IPv6 address. In the case of traffic outside of the
MAP domain, it is translated to the BR's IPv6 prefix. For the BR to
be able to reverse the translation, the full destination IPv4 address
must be encoded in the IPv6 address. The same thing applies if an
IPv4 prefix is encoded in the IPv6 address, then the reverse
translator needs to know the full destination IPv4 address, which has
to be encoded in the interface-id.
There are multiple proposals for how to encode the IPv4 address, and
if also the destinatin port or PSID should also be included. A
couple of the proposals are shown in the figure below.
Note: The encoding of the full IPv4 address into the interface
identifier, both for the source and destination IPv6 addresses have
been shown to be useful for troubleshooting. The format finally
agreed upon here, will apply for both encapsulation and translation.
Existing IANA assigned format [RFC5342]:
| 32 bits | 32 bits |
+------------------+------------------+
| 02-00-5E-FE | IPv4 address |
+------------------+------------------+
Figure 9
Parsable format including the extended IPv4 prefix length (L) and
PSID:
<-8-><-------- L>=32 -------><48-L><8->
+---+----------------+------+-----+---+
| u | IPv4 address | PSID | 0 | L |
+---+----------------+------+-----+---+
Troan Expires May 3, 2012 [Page 18]
Internet-Draft MAP October 2011
Figure 10
If the End-user IPv6 prefix length is larger than 64, the most
significant parts of the interface identifier is overwritten by the
prefix.
Troan Expires May 3, 2012 [Page 19]
Internet-Draft MAP October 2011
6. IANA Considerations
This specification does not require any IANA actions.
Troan Expires May 3, 2012 [Page 20]
Internet-Draft MAP October 2011
7. Security Considerations
There are no new security considerations pertaining to this document.
Troan Expires May 3, 2012 [Page 21]
Internet-Draft MAP October 2011
8. Contributors
The members of the MAP design team are:
Congxiao Bao, Mohamed Boucadair, Gang Chen, Maoke Chen, Wojciech
Dec, Xiaohong Deng, Remi Despres, Jouni Korhonen, Xing Li, Satoru
Matsushima, Tomasz Mrugalski, Tetsuya Murakami, Jacni Qin, Qiong
Sun, Tina Tsou, Dan Wing, Leaf Yeh and Jan Zorz.
Troan Expires May 3, 2012 [Page 22]
Internet-Draft MAP October 2011
9. Acknowledgements
Troan Expires May 3, 2012 [Page 23]
Internet-Draft MAP October 2011
10. References
10.1. Normative References
[I-D.mdt-softwire-map-dhcp-option]
Mrugalski, T., Boucadair, M., and O. Troan, "DHCPv6
Options for Mapping of Address and Port",
draft-mdt-softwire-map-dhcp-option-00 (work in progress),
October 2011.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC5342] Eastlake, D., "IANA Considerations and IETF Protocol Usage
for IEEE 802 Parameters", BCP 141, RFC 5342,
September 2008.
[RFC6346] Bush, R., "The Address plus Port (A+P) Approach to the
IPv4 Address Shortage", RFC 6346, August 2011.
10.2. Informative References
[I-D.bcx-address-fmt-extension]
Bao, C. and X. Li, "Extended IPv6 Addressing for Encoding
Port Range", draft-bcx-address-fmt-extension-02 (work in
progress), October 2011.
[I-D.boucadair-behave-ipv6-portrange]
Boucadair, M., Levis, P., Grimault, J., Villefranque, A.,
Kassi-Lahlou, M., Bajko, G., Lee, Y., Melia, T., and O.
Vautrin, "Flexible IPv6 Migration Scenarios in the Context
of IPv4 Address Shortage",
draft-boucadair-behave-ipv6-portrange-04 (work in
progress), October 2009.
[I-D.boucadair-dhcpv6-shared-address-option]
Boucadair, M., Levis, P., Grimault, J., Savolainen, T.,
and G. Bajko, "Dynamic Host Configuration Protocol
(DHCPv6) Options for Shared IP Addresses Solutions",
draft-boucadair-dhcpv6-shared-address-option-01 (work in
progress), December 2009.
[I-D.boucadair-softwire-stateless-requirements]
Boucadair, M., Bao, C., Skoberne, N., and X. Li,
"Requirements for Extending IPv6 Addressing with Port
Sets", draft-boucadair-softwire-stateless-requirements-00
(work in progress), September 2011.
Troan Expires May 3, 2012 [Page 24]
Internet-Draft MAP October 2011
[I-D.bsd-softwire-stateless-port-index-analysis]
Boucadair, M., Skoberne, N., and W. Dec, "Analysis of Port
Indexing Algorithms",
draft-bsd-softwire-stateless-port-index-analysis-00 (work
in progress), September 2011.
[I-D.chen-softwire-4v6-add-format]
Chen, G. and Z. Cao, "Design Principles of a Unified
Address Format for 4v6",
draft-chen-softwire-4v6-add-format-00 (work in progress),
October 2011.
[I-D.chen-softwire-4v6-pd]
Chen, G., Sun, T., and H. Deng, "Prefix Delegation in
4V6", draft-chen-softwire-4v6-pd-00 (work in progress),
August 2011.
[I-D.dec-stateless-4v6]
Dec, W., Asati, R., Bao, C., Deng, H., and M. Boucadair,
"Stateless 4Via6 Address Sharing",
draft-dec-stateless-4v6-04 (work in progress),
October 2011.
[I-D.despres-softwire-4rd-addmapping]
Despres, R., Qin, J., Perreault, S., and X. Deng,
"Stateless Address Mapping for IPv4 Residual Deployment
(4rd)", draft-despres-softwire-4rd-addmapping-01 (work in
progress), September 2011.
[I-D.despres-softwire-4rd-u]
Despres, R., "Unifying Double Translation and
Encapsulation for 4rd (4rd-U)",
draft-despres-softwire-4rd-u-01 (work in progress),
October 2011.
[I-D.despres-softwire-sam]
Despres, R., "Stateless Address Mapping (SAM) - a
Simplified Mesh-Softwire Model",
draft-despres-softwire-sam-01 (work in progress),
July 2010.
[I-D.despres-softwire-stateless-analysis-tool]
Despres, R., "Analysis of Stateless Solutions for IPv4
Service across IPv6 Networks - A synthetic Analysis Tool",
draft-despres-softwire-stateless-analysis-tool-00 (work in
progress), September 2011.
[I-D.mrugalski-dhc-dhcpv6-4rd]
Troan Expires May 3, 2012 [Page 25]
Internet-Draft MAP October 2011
Mrugalski, T., "DHCPv6 Options for IPv4 Residual
Deployment (4rd)", draft-mrugalski-dhc-dhcpv6-4rd-00 (work
in progress), July 2011.
[I-D.murakami-softwire-4rd]
Murakami, T., Troan, O., and S. Matsushima, "IPv4 Residual
Deployment on IPv6 infrastructure - protocol
specification", draft-murakami-softwire-4rd-01 (work in
progress), September 2011.
[I-D.murakami-softwire-4v6-translation]
Murakami, T., Chen, G., Deng, H., Dec, W., and S.
Matsushima, "4via6 Stateless Translation",
draft-murakami-softwire-4v6-translation-00 (work in
progress), July 2011.
[I-D.sun-softwire-stateless-4over6]
Sun, Q., Xie, C., Cui, Y., Wu, J., Wu, P., Zhou, C., and
Y. Lee, "Stateless 4over6 in access network",
draft-sun-softwire-stateless-4over6-00 (work in progress),
September 2011.
[I-D.xli-behave-divi]
Bao, C., Li, X., Zhai, Y., and W. Shang, "dIVI: Dual-
Stateless IPv4/IPv6 Translation", draft-xli-behave-divi-04
(work in progress), October 2011.
[I-D.xli-behave-divi-pd]
Li, X., Bao, C., Dec, W., Asati, R., Xie, C., and Q. Sun,
"dIVI-pd: Dual-Stateless IPv4/IPv6 Translation with Prefix
Delegation", draft-xli-behave-divi-pd-01 (work in
progress), September 2011.
[RFC1933] Gilligan, R. and E. Nordmark, "Transition Mechanisms for
IPv6 Hosts and Routers", RFC 1933, April 1996.
[RFC2529] Carpenter, B. and C. Jung, "Transmission of IPv6 over IPv4
Domains without Explicit Tunnels", RFC 2529, March 1999.
[RFC2766] Tsirtsis, G. and P. Srisuresh, "Network Address
Translation - Protocol Translation (NAT-PT)", RFC 2766,
February 2000.
[RFC3194] Durand, A. and C. Huitema, "The H-Density Ratio for
Address Assignment Efficiency An Update on the H ratio",
RFC 3194, November 2001.
[RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through
Troan Expires May 3, 2012 [Page 26]
Internet-Draft MAP October 2011
Network Address Translations (NATs)", RFC 4380,
February 2006.
[RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site
Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214,
March 2008.
[RFC5969] Townsley, W. and O. Troan, "IPv6 Rapid Deployment on IPv4
Infrastructures (6rd) -- Protocol Specification",
RFC 5969, August 2010.
[RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X.
Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052,
October 2010.
[RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful
NAT64: Network Address and Protocol Translation from IPv6
Clients to IPv4 Servers", RFC 6146, April 2011.
[RFC6147] Bagnulo, M., Sullivan, A., Matthews, P., and I. van
Beijnum, "DNS64: DNS Extensions for Network Address
Translation from IPv6 Clients to IPv4 Servers", RFC 6147,
April 2011.
[RFC6333] Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual-
Stack Lite Broadband Deployments Following IPv4
Exhaustion", RFC 6333, August 2011.
Troan Expires May 3, 2012 [Page 27]
Internet-Draft MAP October 2011
Appendix A. Open issues / New features
A.1. Max PSID
It has been proposed to keep independence of IPv6 routing plans from
IPv4 considerations and yet to be able to support variable sized port
sets per shared IPv4 address. A mechanism proposed for this is
called "MAX PSID". The idea is that a source, transmitting a packet
to a CE doesn't need to know the length of the PSID field of that CE.
All port bits after offset bits are copied in the encoded IPv6
address. This implies that a MAP CE be capable of receiving MAP
traffic for multiple addresses within its delegated prefix, e.g.
using the same mechanism as used for double translation when CEs are
allocated IPv4 prefixes shorter than /32.
A.2. Interface identifier - V octet and Checksum neutrality
There are multiple issues related to the Interface-identifier
encoding.
o The V octet is required to distinguish between MAP and native IPv6
traffic if the same End-user IPv6 prefix is used. If a separate
End-user IPv6 prefix is used for MAP traffic, requiring a special
flag in the interface-identifier is not required.
o The Checksum-neutrality preserver (CNP). It is for MAP packets to
be acceptable by IPv6 functions that check UDP/TCP checksums,
without needing for this to consider transport-layer fields.
Checksum neutrality is useful for double translation and, possibly
more important, it permits to envisage a unified solution which
has significant advantages of both encapsulation and double
translation [I-D.despres-softwire-4rd-u]. With encapsulation, the
field can be set to 0.
With both mechanisms, IPv6 addresses have the following format:
|<--------------- 64 ------------><8><----- 40 ------><--16--->
+---------------------------------+-+----------------+--------+
| Unformatted IPv6 prefix (part 1)|V| (part 2) |CNP or 0|
+---------------------------------+-+----------------+--------+
The V octet deterministically differentiates MAP addresses from other
IPv6 addresses by having its bits 6 and 7 set to 1 and 1 (they are 1
and 0 in modified EUI-64 Interface-ID format, and bit 6 is 0 in the
privacy extension of [RFC 3041]. V is proposed to be 0x03 (which
leaves 2^6 values of bits 0 to 5 for other Interface ID formats that
Troan Expires May 3, 2012 [Page 28]
Internet-Draft MAP October 2011
could be useful in the future).
The Unformatted IPv6 prefix starts with bits derived from the IPv4
address being mapped (e.g. Rule IPv6 prefix, IPv4 suffix, and PSID,
or Max PSID if applicable). The remainder to reach 104 bits is
filled with 0s.
The CNP field is, in one's complement arithmetic, the sum of the two
halves of the IPv4 address, minus the sum of the seven 16-bit fields
that precede the CNP in the IPv6 address.
A.3. Optional BR per Rule within a domain
With BR IPv6 address/prefix as optional parameters in mapping rules,
it has been proposed to support ISP networks that have IPv4 prefixes
coming from several providers necessitating geographically dispersed
BRs. In such configurations, each provider exercises ingress
filtering so that a CE MUST sent its traffic going to the Internet
via the right BR (that whose locally routed IPv4 prefixes include one
that matches the IPv4 address or prefix of the CE)
[I-D.despres-softwire-4rd-addmapping].
Troan Expires May 3, 2012 [Page 29]
Internet-Draft MAP October 2011
Appendix B. Requirements
This list of requirements for a stateless mapping of address and
ports solution may not be complete. The requirements are listed in
no particular order, and they may be conflicting.
R-1: To allow for a single user delegated IPv6 prefix to be used
for native IPv6 service and for MAP, the representation of an
IPv4 prefix, address or shared IPv4 address and PSID must be
efficient. As an example it must be possible to represent a
shared IPv4 address and PSID in 24 bits or less. (Given a
typical prefix assignment of /56 to the end-user and a MAP
IPv6 prefix of /32.)
R-2: The IPv6 address format and mapping must be flexible, and
support any placement of the embedded bits from IPv4 prefix/
address and port set in the IPv6 address.
R-3: Algorithm complexity. The mapping from an IPv4 address and
port to an IPv6 address is done in the forwarding plane on MAP
nodes. It is important that the algorithm is bounded and as
efficient as possible.
R-4: MAP must allow service providers to define their own address
sharing ratio. MAP MUST NOT in particular restrict by design
the possible address sharing ratio; ideally 1:1 and 1:65536
should be supported. The mapping must at least support a
sharing ratio of 64, 1024 ports per end-user.
R-5: The mapping may support deployments using differentiated port-
sets. That is, end-users are assigned different sized port-
sets and direct communication between MAP CEs are permitted.
R-6: The mapping should support differentiated port sets within a
single shared IPv4 address. (i.e., be able to assign port sets
of different sizes to customers without requiring any per
customer state to be instantiated in network elements involved
in data transfer).
R-7: The MAP solution should support excluding the well known ports
0-1023.
R-8: It MUST be possible to assign well known ports to a CE.
R-9: There must not be any dependency between IPv6 addressing and
IPv4 addressing. With the exception where full IPv4 addresses
or prefixes are encoded. Then IPv6 prefix assignment must be
done so that martian IPv4 addresses are not assigned.
Troan Expires May 3, 2012 [Page 30]
Internet-Draft MAP October 2011
R-10: The mapping must not require IPv4 routing to be imported in
IPv6 routing.
R-11: The mapping should support legacy RTP/RTCP compatibility.
(Allocating two consecutive ports).
R-12: The mapping may be UPnP 1.0 friendly. A UPnP client will keep
asking for the next port (as in current port + 1) a scattered
port allocation will be more UPnP friendly.
R-13: For out of domain traffic the mapping must support embedding a
full IPv4 address in the IPv6 interface identifier. This is
required in the translation case. It also simplifies pretty
printing and other operational tools.
R-14: For Service Providers requiring to apply specific policies on
per Address-Family (e.g., IPv4, IPv6), some provisioning tools
(e.g., DHCPv6 option) may be required to derive in a
deterministic way the IPv6 address to be used for the IPv4
traffic based on the IPv6 prefix delegated to the home
network.
R-15: It should/must/may be possible to use the same IPv6 prefix
(/64) for native IPv6 traffic and MAPed traffic.
R-16: When only one single IPv6 prefix is assigned for both native
IPv6 communications and the transport of IPv4 packets, the
IPv4-translatable IPv6 prefix MUST have a length less than
/64. When distinct prefixes are used, this requirement is
relaxed.
R-17: The same mapping must support both translation and
encapsulation solutions.
Troan Expires May 3, 2012 [Page 31]
Internet-Draft MAP October 2011
Appendix C. Deployment considerations
C.1. Flexible Assigment of Port Sets
Different classes of customers require port sets of different size.
In the context of shared IPv4 addresses, some customers would be
satisfied with an shared IPv4 addresses, while others may need to be
assigned a single IPv4 address or delegated an IPv4 subnet.
C.2. Traffic Classification
Usually, ISPs adopt traffic classification to ensure service quality
for different classes of customers. This feature is also helpful for
customer behavior monitoring and security protection. For example,
DIA (Dedicated Internet Access) has been provided by operators for
corporations to cater for their Internet communications needs.
Service is made by means of the edge router features and key systems,
like ACL (Access List Control) to classify different traffic. Five
tuples would be identified from IP header and UDP/TCP header.
Currently, it is very well supported in IPv4. Vendors are delivering
or committed to support that feature for IPv6. In order to
facilitating IPv6 deployment, MAP solution should support this
feature on IPv6 plane.
C.3. Prefix Delegation Deployment
Prefix delegation is an important feature both for broadband and
mobile network. In mobile network, prefix delegation is introduced
in 3GPP network in Release 10. The deployment of PD would be
supported in 4v6 case. Variable length of IPv6 prefix is assigned to
CPE for deriving IPv4 information.
C.4. Coexisting Deployment
4v6 solutions(i.e. encapsulation and translation) would not only
coexist with each other, but also can harmonize with other deployment
cases. Here lists some coexisting cases. (Note: more coexisting
cases are expected to be investigated in future.)
o Case 1: Coexisting between 4v6 encapsulation and 4v6 translation
o Case 2: Coexisting between 4v6 translation and NAT64 (Single
Translation)
o Case 3: Coexisting between 4v6 solutions and SLAAC
Troan Expires May 3, 2012 [Page 32]
Internet-Draft MAP October 2011
C.5. Friendly to Network Provisioning
Network management plane normally has an ability to to identify
different users and the compatible with the address assignment
techniques in the domain. 4v6 would conform to current practices on
management plane. In 3GPP network, for example, only the IPv6 prefix
is assigned to the devices, used to identify different users. And
management plane for one family address is better than two, namely
the operating platform does not need to manage both IPv4 and IPv6.
Since only IPv6 prefix is assigned, 4v6 on the management plane is
naturally conducted only via IPv6.
C.6. Enable privacy addresses
User privacy should be taken into account when 4v6 solution is
deployed. Some security concern associated with non-changing IPv6
interface identifiers has been expressed in RFC4941[RFC4941].
Ability to change the interface identifier over time makes it more
difficult for eavesdroppers and other information collectors to
identify when different addresses used in different sessions actually
correspond to the same node.
C.7. Facilitating 4v6 Service
Some ISPs may need to offer services in a 4v6 domain with a shared
address, e.g. 4v6 node hosts FTP server. The service provisioning
may require well-know port range(i.e. port range belong to 0-1023).
MAP would provide operators with possibilities to generate a port
range including the 0-1023. Afterwards, operators could decide to
assign it to any requesting user.
C.8. Independency with IPv6 Routing Planning
The IPv6 routing is easier to plan if it's not impacted by the
encoded IPv4 or port ID information. MAP would prohibit IPv4 routing
imported in IPv6.
C.9. Optimized Routing Path
MAP could achieve optimized routing path both for hub case and mesh
case. Traffic in hub and spoke case could follow asymmetric routing,
in which incoming routes would not be binded to a given border point
but others geographically closed to traffic initiators. In mesh
cases, traffic between CPEs could directly communicate without going
through remote border point.
Troan Expires May 3, 2012 [Page 33]
Internet-Draft MAP October 2011
Appendix D. Guidelines for Operators
This appendix is purposed to (1) clarify the difference and relevance
of MAP address mapping format and what has published in standard
track; (2) provide a referential guideline to operators, illustrating
a common use-case of MAP deployment.
D.1. Additional terms
The following terms are listed, mainly used in this appendix only, as
an add-on to the terminology of the main text.
4pfx the index for an IPv4 prefix, either generated
with coding or as same as the IPv4 prefix
itslef.
ug-octet the octet consisting of 64-71 bits in the IPv6
address, containing the bits u and g defined by
EUI-64 standard.
Common prefix an aggregate decided by a domain for the MAP
deployment. It is a subset of the operator's
aggregates by its RIR or provider.
IPv4 suffix the part of IPv4 address bits used for
identifying CEs.
Host suffix the IPv6 suffix assigning to an end system.
NOTE: it doesn't mean this should be really
configured on a certain interface of a host.
MAP-format the address mapping format defined by this
document.
RFC6052-format the address mapping format defined by [RFC6052]
and its succeeding extensions (or updates) for
port-space sharing, for example,
[I-D.xli-behave-divi].
D.2. Understanding address formats: their difference and relevance
MAP introduces an address format of embedding IPv4 information to
IPv6 address. On the other hand, we also have [RFC6052] defines an
address format with the similar property. With extending port-set
id, it can also support address sharing among different CEs
[I-D.xli-behave-divi]. What are their difference and relevance?
We present a common abstract format for them both, as is depicted in
Troan Expires May 3, 2012 [Page 34]
Internet-Draft MAP October 2011
Figure 11. For the easy expression, we exclude the ug-octet, which
is not concerned in this appendix.
|<----- 120 bits (IPv6 address excluding ug-octet) --------->|
+-------------+------+-------------+------+-----------//-----+
|Common Prefix| 4pfx | IPv4 suffix | PSID | Host Suffix |
+-------------+------+-------------+------+-----------//-----+
Figure 11: Abstract view of MAP- and RFC6052-formats
Only two parts in Figure 11 are different for MAP- and RFC6052-
formats. We compare them in Figure 12 and following paragraphs.
+----------------+--------------+------------+
| | MAP | RFC6052 |
+----------------+--------------+------------+
| from IPv4 | coding with | same, w/o |
| prefix to 4pfx | compression | change |
+----------------+--------------+------------+
| Host | full v4.addr | padding to |
| Suffix | or 4rd IID | zero |
+----------------+--------------+------------+
Figure 12: Difference between MAP- and RFC6052-formats
The comparison clarifies that the major role of full IPv4 address
embedded in the RFC6052 format is replaced by the MAP's coded IPv4
prefix index and the uncoded IPv4 suffix. The Figure 13 illustrates
this relevance.
Troan Expires May 3, 2012 [Page 35]
Internet-Draft MAP October 2011
(delegated prefix in RFC6052 format, w/o rule)
+-------------+-------------+-------------+------+
|Common Prefix| full IPv4 address (32bit) | PSID |
+-------------+-------------+-------------+------+
: : : :
+-------------+-------------+
32 bits: | 4pfx | IPv4 suffix | :
+-------------+-------------+ +
: . . .
: . . .
: . . .
+-----+-------------+ +
m bits: |4pfx | IPv4 suffix | : (w/ rules)
+-----+-------------+
: : : :
+-------------+-----+-------------+------+
| Rule IPv6 Prefix | CE index |
+-------------+-----+-------------+------+
(delegated prefix in MAP format)
Figure 13: Relevance between MAP- and RFC6052-formats
o Why is it needed to compress the IPv4 prefix?
Precisely speaking, it is not "to compress the IPv4 prefix" but
"to establish correspondence between IPv6 delegated prefixes and
the residual IPv4 prefixes."
It is important for an operator to understand what the MAP is
designed for and where it could be applied. A keyword for MAP is
"residual deployment", referring to the deployment of an IPv6
network with utilizing the residual IPv4 address spaces for the
subnets/host having IPv4 communication, without introducing per-
session states.
Therefore, the delegated CE prefixes are determined prior to
finding a proper IPv4 address block in hand to be mapped to the CE
index and the IPv4 prefix index (4pfx) as well as the Rule IPv6
prefix.
IPv6 delegation planning, independent of the IPv4 addressing, also
implies to follow the common convention of assigning a /64 prefix
to any IPv6 local network. It is highly impossible to directly
match some IPv4 prefixes to the already-determined IPv6 prefixes,
Troan Expires May 3, 2012 [Page 36]
Internet-Draft MAP October 2011
and therefore the prefixes have to be coded and typically it is a
compression.
If we have a short-enough Common Prefix, it is also possible to
deploy a direct matching where 4pfx is equal to IPv4 prefix. Only
in this case, the MAP-format is equivalent to the RFC6052-format
and the rule set could be simplified to a unique rule for
0.0.0.0/0.
Once the unique rule for 0.0.0.0/0 is defined, the special rule
for the out-of-domain traffic towards the BR is not needed any
more. The route with the common prefix itself can play the role
of less-specific routes for the whole IPv4 space. This is a
feature of the RFC6052-format.
o Why does MAP copy IPv4 address in the suffix?
The full IPv4 address is copied in the Host suffix of MAP with two
reasons.
First of all, for the traffic going out of the domain, compress
coding makes full IPv4 address information not directly appear in
the IPv6 prefix for BR at all. To enable the double translation,
it is had to embed this information in the Host Suffix of MAP-
format for the peer IPv4 address outside of the domain.
Further, it is not necessary to separate the processing for the
in-domain addresses and that for the out-domain addresses. Making
a symmetric format is perferred.
Another concern is the simplicity. Even though the delegated
prefix is theorectally sufficient to extract the corresponding
IPv4 address for the CE, it relies on retrieving rules for every
datagram. Embedding the full IPv4 address in the suffix
simplifies the processing at IPv6-to-IPv4 translator when
utilizing MAP for double translation. It also helps in setting
filters at middle boxes, with exposing the IPv4 full addresses
dispatched to the CEs.
MAP is designed for the residual deployment, including the case of
recalling deployed IPv4 addresses and reallocating them for the
deployment in IPv6 networks. To this extent, MAP can understood as
"4rd-MAP".
Troan Expires May 3, 2012 [Page 37]
Internet-Draft MAP October 2011
In practice, MAP-format can be also used for the objective of
providing stateless encapsulation or double translation for the
already deployed IPv4 networks, without renumbering, whose provider
backbone is upgraded to IPv6. Unlike the residual deployment, this
use-case unavoidably introduces IPv4 routing entropy into the IPv6
routing infrastrucutre. On the other hand, for the old IPv4 network
or IPv6 network upgraded from IPv4, it is not necessarily having 64
bits for their host identifiers. Therefore longer-than-/64 prefix is
not a strict constrain. Therefore, RFC6052-format is recommended in
this case of non-residual deployment. RFC6052-format is motivated
with keeping temporal uniqueness of end-to-end identifiers throughout
the period of transition and providing the rule-free simplicity.
D.3. Residual deployment with MAP
This section illustrate how we can use MAP in the operation of
residual deployment.
NOTE: Applying MAP for a use-case other than residual deployment
should follow different logic of address planning and therefore,
because of the reason mentioned above, not included in this Appendix.
Residual deployment starts from IPv6 address planning. A simple
example is taken inline for easy understanding.
(A) IPv6 considerations
(A1) Determine the maximum number N of CEs to be supported, and, for
generality, suppose N = 2^n.
For example, we suppose n = 20. It means there will be up to
about one million CEs.
(A2) Choose the length x of IPv6 prefixes to be assigned to ordinary
customers.
Considering we have a /32 IPv6 block, it is not a problem for
the IPv6 deployment with the given number of CEs. Let x = 60,
allowing subnets inside in each CE delegated networks.
Troan Expires May 3, 2012 [Page 38]
Internet-Draft MAP October 2011
(A3) Multiply N by a margin coefficient K, a power of two (K = 2 ^
k), to take into account that:
- Some privileged customers may be assigned IPv6 prefixes of
length x', shorter than x, to have larger addressing spaces
than ordinary customers, both in IPv6 and IPv4;
- Due to the hierarchy of routable prefixes, many theoretically
delegatable prefixes may not be actually delegatable (ref: host
density ratio of [RFC3194]).
In our example, let's take k = 0 for simplicity.
(B) IPv4 considerations
(B1) List all (non overlapping, not yet assigned to any in-running
networks) IPv4 prefixes Hi that are available for IPv4 residual
deployment.
Suppose that we hold two blocks and not yet assigned to any
fixed network: 192.32../16 and 63.245../16.
(B2) Take enough of them, among the shortest ones, to get a total
whose size M is a power of two (M = 2 ^ m), and includes a good
proportion of the available IPv4 space.
If the M < N, addresses should be shared by N CEs and thus each
is shared by N/M = 2^(n - m) CEs with PSID length of (n - m).
If we use both blocks, M = 2^16 + 2^16, and therefore m = 17.
Then PSID length could be 3 bits, the corresponding sharing
ratio is also determined so that each CE can have 8192 ports to
use under the shared global IPv4 address.
Troan Expires May 3, 2012 [Page 39]
Internet-Draft MAP October 2011
(B3) For each IPv4 prefix Hi of length hi, choose a "rule index",
i.e., the 4pfx in Fig.C-1 and Fig.C-3, say Ri of length ri = m
- (32 - hi).
All these indexes must be non overlapping prefixes (e.g. 0, 10,
110, 111 for one /10, one /11, and two /12).
Then we have:
H1 = 192.32../16, h1 = 16, r1 = 1 => R1 = bin(0);
H2 = 63.245../16, h2 = 16, r2 = 1 => R2 = bin(1);
(C) After (A) and (B), derive the rule(s)
(C1) Derive the length c of the "Common prefix" C that will appear
at the beginning of all delegated prefixes (c = x - (n + k)).
(C2) Take any prefix for this C of length c that starts with a RIR-
allocated IPv6 prefix.
(C3) For each IPv4 prefix Hi, make a rule, in which the key is Hi,
and the value is the Common prefix C followed by the Rule index
Ri. Then this i-th rule's Rule IPv6 Prefix will have the
length of (c + ri).
Then we can do that:
c = 40 => C = 2001:0db8:ff00::/40
Rule 1: Rule IPv6 Prefix = 2001:0db8:ff00::/41
Rule 2: Rule IPv6 Prefix = 2001:0db8:ff80::/41
Troan Expires May 3, 2012 [Page 40]
Internet-Draft MAP October 2011
As a result, for a certain CE delegating 2001:0db8:ff98:
7650::/60, its parameters are:
Rule IPv6 Prefix = 2001:0db8:ff80::/41 => Rule 2
IPv4 Suffix = bin(001 1000 0111 0110 0)
PSID = bin(101) = 0x5
Rule IPv4 Prefix = 63.245../16
CE IPv4 Address = 63.245.48.236
If different sharing ratio is expected, we may partition CEs into
groups and do (A) and (B) for each group, determining the PSID length
for them separately. However, this might cause a fairly complicated
work in the address planning.
Troan Expires May 3, 2012 [Page 41]
Internet-Draft MAP October 2011
Author's Address
Ole Troan (editor)
cisco
Oslo
Norway
Email: ot@cisco.com
Troan Expires May 3, 2012 [Page 42]