One document matched: draft-ietf-dhc-sedhcpv6-06.xml
<?xml version="1.0" encoding="US-ASCII"?>
<!-- This template is for creating an Internet Draft using xml2rfc,
which is available here: http://xml.resource.org. -->
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
<!-- One method to get references from the online citation libraries.
There has to be one entity for each item to be referenced.
An alternate method (rfc include) is described in the references. -->
<!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml">
<!ENTITY RFC2629 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2629.xml">
]>
<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>
<!-- used by XSLT processors -->
<!-- For a complete list and description of processing instructions (PIs),
please see http://xml.resource.org/authoring/README.html. -->
<?rfc strict="yes" ?>
<!-- give errors regarding ID-nits and DTD validation -->
<!-- control the table of contents (ToC) -->
<?rfc toc="yes"?>
<!-- generate a ToC -->
<?rfc tocdepth="3"?>
<!-- the number of levels of subsections in ToC. default: 3 -->
<!-- control references -->
<?rfc symrefs="yes"?>
<!-- use symbolic references tags, i.e, [RFC2119] instead of [1] -->
<?rfc sortrefs="yes" ?>
<!-- sort the reference entries alphabetically -->
<!-- control vertical white space
(using these PIs as follows is recommended by the RFC Editor) -->
<?rfc compact="yes" ?>
<!-- do not start each main section on a new page -->
<?rfc subcompact="no" ?>
<!-- keep one blank line between list items -->
<!-- end of list of popular I-D processing instructions -->
<rfc category="std" docName="draft-ietf-dhc-sedhcpv6-06" ipr="trust200902">
<front>
<title abbrev="SeDHCPv6">Secure DHCPv6</title>
<author fullname="Sheng Jiang" initials="S." role="editor" surname="Jiang">
<organization>Huawei Technologies Co., Ltd</organization>
<address>
<postal>
<street>Q14, Huawei Campus, No.156 Beiqing Road</street>
<city>Hai-Dian District, Beijing, 100095</city>
<country>CN</country>
</postal>
<email>jiangsheng@huawei.com</email>
</address>
</author>
<author fullname="Sean Shen" initials="S." surname="Shen">
<organization>CNNIC</organization>
<address>
<postal>
<street>4, South 4th Street, Zhongguancun</street>
<city>Beijing</city>
<code>100190</code>
<country>CN</country>
</postal>
<email>shenshuo@cnnic.cn</email>
</address>
</author>
<author fullname="Dacheng Zhang" initials="D." surname="Zhang">
<organization>Huawei Technologies Co., Ltd</organization>
<address>
<postal>
<street>Q14, Huawei Campus, No.156 Beiqing Road</street>
<city>Hai-Dian District, Beijing, 100095</city>
<country>CN</country>
</postal>
<email>zhangdacheng@huawei.com</email>
</address>
</author>
<author fullname="Tatuya Jinmei" initials="T." surname="Jinmei">
<organization>Infoblox Inc.</organization>
<address>
<postal>
<street>3111 Coronado Drive</street>
<city>Santa Clara</city>
<region>CA</region>
<country>US</country>
</postal>
<email>jinmei@wide.ad.jp</email>
</address>
</author>
<date month="" year="2015" />
<area>Internet Area</area>
<workgroup>DHC Working Group</workgroup>
<keyword>Secure</keyword>
<keyword>DHCPv6</keyword>
<keyword>Public Key</keyword>
<abstract>
<t>The Dynamic Host Configuration Protocol for IPv6 (DHCPv6) enables
DHCPv6 servers to pass configuration parameters. It offers configuration
flexibility. If not being secured, DHCPv6 is vulnerable to various
attacks, particularly spoofing attacks. This document analyzes the
security issues of DHCPv6 and specifies a Secure DHCPv6 mechanism for
communications between DHCPv6 clients and DHCPv6 servers. This document
provides a DHCPv6 client/server authentication mechanism based on
sender's public/private key pairs or certificates with associated
private keys. The DHCPv6 message exchanges are protected by the
signature option and the timestamp option newly defined in this
document.</t>
</abstract>
</front>
<middle>
<section title="Introduction">
<t>The Dynamic Host Configuration Protocol for IPv6 (DHCPv6, <xref
target="RFC3315"></xref>) enables DHCPv6 servers to pass configuration
parameters and offers configuration flexibility. If not being secured,
DHCPv6 is vulnerable to various attacks, particularly spoofing
attacks.</t>
<t>This document analyzes the security issues of DHCPv6 in details. This
document provides mechanisms for improving the security of DHCPv6
between client and server:<list style="symbols">
<t>the identity of a DHCPv6 message sender, which can be a DHCPv6
server or a client, can be verified by a recipient.</t>
<t>the integrity of DHCPv6 messages can be checked by the recipient
of the message.</t>
<t>anti-replay protection based on timestamps.</t>
</list></t>
<t>Note: this secure mechanism in this document does not protect the
relay-relevant options, either added by a relay agent toward a server or
added by a server toward a relay agent,
because they are only transported within operator networks
and considered less vulnerable.
Communication between a server and a relay agent, and communications
between relay agents, may be secured through the use of IPsec, as
described in section 21.1 in <xref target="RFC3315"></xref>.</t>
<t>The security mechanisms specified in this document is based on
sender's public/private key pairs or certificates with associated
private keys. The reason for such design and deployment consideration
are discussed in <xref target="DeployConsider"></xref>. It also
integrates message signatures for the integrity and timestamps for
anti-replay. The sender authentication procedure using certificates
defined in this document depends on deployed Public Key Infrastructure
(PKI, <xref target="RFC5280"></xref>). However, the deployment of PKI is
out of the scope.</t>
<t>Secure DHCPv6 is applicable in environments where physical security
on the link is not assured (such as over wireless) and attacks on DHCPv6
are a concern.</t>
</section>
<section title="Requirements Language and Terminology">
<t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in <xref
target="RFC2119"></xref> when they appear in ALL CAPS. When these words
are not in ALL CAPS (such as "should" or "Should"), they have their
usual English meanings, and are not to be interpreted as <xref
target="RFC2119"></xref> key words.</t>
</section>
<section title="Security Overview of DHCPv6">
<t>DHCPv6 is a client/server protocol that provides managed
configuration of devices. It enables a DHCPv6 server to automatically
configure relevant network parameters on clients. In the basic DHCPv6
specification <xref target="RFC3315"></xref>, security of DHCPv6
messages can be improved.</t>
<t>The basic DHCPv6 specifications can optionally authenticate the
origin of messages and validate the integrity of messages using an
authentication option with a symmetric key pair. <xref
target="RFC3315"></xref> relies on pre-established secret keys. For any
kind of meaningful security, each DHCPv6 client would need to be
configured with its own secret key; <xref target="RFC3315"></xref>
provides no mechanism for doing this.</t>
<t>For the keyed hash function, there are two key management mechanisms.
The first one is a key management done out of band, usually through some
manual process. The second approach is to use Public Key Infrastructure
(PKI).</t>
<t>As an example of the first approach, operators can set up a key
database for both servers and clients from which the client obtains a key
before running DHCPv6. Manual key distribution runs counter to the goal
of minimizing the configuration data needed at each host.</t>
<t><xref target="RFC3315"></xref> provides an additional mechanism for
preventing off-network timing attacks using the Reconfigure message: the
Reconfigure Key authentication method. However, this method provides
little message integrity or source integrity check, and it protects only
the Reconfigure message. This key is transmitted in plaintext.</t>
<t>In comparison, the security mechanism defined in this document allows
the public key database on the client or server to be populated
opportunistically or manually, depending on the degree of confidence
desired in a specific application. PKI security mechanism is simpler in
the local key management respect.</t>
</section>
<section title="Overview of Secure DHCPv6 Mechanism with Public Key">
<t>This document introduces a Secure DHCPv6 mechanism that uses
signatures to secure the DHCPv6 protocol. In order to enable DHCPv6
clients and servers to perform mutual authentication without previous
key deployment, this solution provides a DHCPv6 client/server
authentication mechanism based on public/private key pairs and,
optionally, PKI certificates. The purpose of this design
is to make it easier to deploy DHCPv6 authentication and provides
protection of DHCPv6 message within the scope of whatever trust
relationship exists for the particular key used to authenticate the
message.</t>
<t>In this document, we introduce a public key option, a certificate
option, a signature option and a timestamp option with corresponding
verification mechanisms. A DHCPv6 message can include a public key
option, and carrying a digital signature and a timestamp option. The
signature can be verified using the supplied public key. The recipient
processes the payload of the DHCPv6 message only if the validation is
successful: the signature validates, and a trust relationship exists for
the key. Alternatively, a DHCPv6 message can include a certificate
option, and also carrying a digital signature and a timestamp option.
The signature can be verified by the recipient. The recipient processes
the payload of the DHCPv6 message only if the validation is successful:
the certificate validates, and a trust relationship exists on the
recipient for the provided certificate. The recipient processes the
payload of the DHCPv6 message only if the validation is successful. The
end-to-end security protection can be bidirectional, covering messages
from servers to clients and from clients to servers. Additionally, the
optional timestamp mechanism provides anti-replay protection.</t>
<t>A trust relationship for a public key can be the result either of a
Trust-on-first-use (TOFU) policy, or a list of trusted keys configured
on the recipient.</t>
<t>A trust relationship for a certificate could also be treated either
as Trust-on-first-use or configured in a list of trusted certificate
authorities, depending on the application. Such applications are out of
scope for this document. </t>
<t>Secure DHCPv6 messages are commonly large. One example is normal
DHCPv6 message length plus a 1 KB for a X.509 certificate and signature
and 256 Byte for a signature. IPv6 fragments <xref
target="RFC2460"></xref> are highly possible. In practise, the total
length would be various in a large range. Hence, deployment of Secure
DHCPv6 should also consider the issues of IP fragment, PMTU, etc. Also,
if there are firewalls between secure DHCPv6 clients and secure DHCPv6
servers, it is RECOMMENDED that the firewalls are configured to pass
ICMP Packet Too Big messages <xref target="RFC4443"></xref>.</t>
<section title="New Components">
<t>The components of the solution specified in this document are as
follows:</t>
<t><list style="symbols">
<t>Servers and clients using public keys in their secure DHCPv6
messages generate a public/private key pair. A DHCPv6 option that
carries the public key is defined.</t>
<t>Servers and clients that use certifiicates first generate a
public/private key pair and then obtain a public key certificate
from a Certificate Authority that signs the public key. Another
option is defined to carry the certificate.</t>
<t>A signature generated using the private key which is used by
the receiver to verify the integrity of the DHCPv6 messages and
then the identity of the sender.</t>
<t>A timestamp, to detect replayed packet. The secure DHCPv6 nodes
need to meet some accuracy requirements and be synced to global
time, while the timestamp checking mechanism allows a configurable
time value for clock drift. The real time provision is out of
scope of this document.</t>
</list></t>
</section>
<section title="Support for Algorithm Agility">
<t>Hash functions are used to provide message integrity checks. In
order to provide a means of addressing problems that may emerge in the
future with existing hash algorithms, as recommended in <xref
target="RFC4270"></xref>, this document provides a mechanism for
negotiating the use of more secure hashes in the future.</t>
<t>In addition to hash algorithm agility, this document also provides
a mechanism for signature algorithm agility.</t>
<t>The support for algorithm agility in this document is mainly a
unilateral notification mechanism from sender to recipient. A
recipient MAY support various algorithms simultaneously among
different senders, and the different senders in a same administrative
domain may be allowed to use various algorithms simultaneously. It is
NOT RECOMMENDED that the same sender and recipient use various
algorithms in a single communication session.</t>
<t>If the recipient does not support the algorithm used by the sender,
it cannot authenticate the message. In the client-to-server case, the
server SHOULD reply with an AlgorithmNotSupported status code (defined
in <xref target="StatusCodes"></xref>). Upon receiving this status
code, the client MAY resend the message protected with the mandatory
algorithm (defined in <xref target="SigOption"></xref>).</t>
</section>
<section title="Applicability">
<t>By default, a secure DHCPv6 enabled client or server SHOULD start
with secure mode by sending secure DHCPv6 messages. If the recipient
is secure DHCPv6 enabled and the key or certificate authority is
trusted by the recipient, then their communication would be in secure
mode. In the scenario where the secure DHCPv6 enabled client and
server fail to build up secure communication between them, the secure
DHCPv6 enabled client MAY choose to send unsecured DHCPv6 message
towards the server according to its local policies.</t>
<t>In the scenario where the recipient is a legacy DHCPv6 server that
does not support secure mechanism, the DHCPv6 server (for all of known
DHCPv6 implementations) would just omit or disregard unknown options
(secure options defined in this document) and still process the known
options. The reply message would be unsecured, of course. It is up to
the local policy of the client whether to accept the messages. If the
client accepts the unsecured messages from the DHCPv6 server, the
subsequent exchanges will be in the unsecured mode.</t>
<t>In the scenario where a legacy client sends an unsecured message to
a secure DHCPv6 enabled server, there are two possibilities depending
on the server policy. If the server's policy requires the
authentication, an UnspecFail (value 1, <xref
target="RFC3315"></xref>) error status code, SHOULD be returned. In
such case, the client cannot build up the connection with the server.
If the server has been configured to support unsecured clients, the
server MAY fall back to the unsecured DHCPv6 mode, and reply unsecured
messages toward the client; depending on the local policy, the server
MAY continue to send the secured reply messages with the consumption
of computing resource. The resources allocated for unsecured clients
SHOULD be separated and restricted.</t>
</section>
</section>
<section title="Extensions for Secure DHCPv6">
<t>This section describes the extensions to DHCPv6. Four new options
have been defined. The new options MUST be supported in the Secure
DHCPv6 message exchange.</t>
<section anchor="PKOption" title="Public Key Option">
<t>The Public Key option carries the public key of the sender. The
format of the Public Key option is described as follows:</t>
<t><figure align="center">
<artwork><![CDATA[ 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OPTION_PUBLIC_KEY | option-len |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
. Public Key (variable length) .
. .
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
option-code OPTION_PUBLIC_KEY (TBA1).
option-len Length of public key in octets.
Public Key A variable-length field containing a
SubjectPublicKeyInfo object specified in [RFC5280].
The SubjectPublicKeyInfo structure is comprised with
a public key and an AlgorithmIdentifier object
which is specified in section 4.1.1.2, [RFC5280]. The
object identifiers for the supported algorithms and
the methods for encoding the public key materials
(public key and parameters) are specified in
[RFC3279], [RFC4055], and [RFC4491].
]]></artwork>
</figure></t>
</section>
<section anchor="CertOption" title="Certificate Option">
<t>The Certificate option carries the public key certificate of the
client. The format of the Certificate option is described as
follows:</t>
<t><figure align="center">
<artwork><![CDATA[ 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OPTION_CERTIFICATE | option-len |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
. Certificate (variable length) .
. .
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
option-code OPTION_CERTIFICATE (TBA2).
option-len Length of certificate in octets.
Certificate A variable-length field containing certificate. The
encoding of certificate and certificate data MUST
be in format as defined in Section 3.6, [RFC7296].
The support of X.509 certificate - Signature (4)
is mandatory.
]]></artwork>
</figure></t>
</section>
<section anchor="SigOption" title="Signature Option">
<t>The Signature option allows a signature that is signed by the
private key to be attached to a DHCPv6 message. The Signature option
could be any place within the DHCPv6 message while it is logically
created after the entire DHCPv6 header and options, except for the
Authentication Option. It protects the entire DHCPv6 header and
options, including itself, except for the Authentication Option. The
format of the Signature option is described as follows:</t>
<t><figure align="center">
<artwork><![CDATA[ 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OPTION_SIGNATURE | option-len |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| HA-id | SA-id | |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| |
. Signature (variable length) .
. .
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
option-code OPTION_SIGNATURE (TBA3).
option-len 2 + Length of Signature field in octets.
HA-id Hash Algorithm id. The hash algorithm is used for
computing the signature result. This design is
adopted in order to provide hash algorithm agility.
The value is from the Hash Algorithm for Secure
DHCPv6 registry in IANA. The support of SHA-256 is
mandatory. A registry of the initial assigned values
is defined in Section 8.
SA-id Signature Algorithm id. The signature algorithm is
used for computing the signature result. This
design is adopted in order to provide signature
algorithm agility. The value is from the Signature
Algorithm for Secure DHCPv6 registry in IANA. The
support of RSASSA-PKCS1-v1_5 is mandatory. A
registry of the initial assigned values is defined
in Section 8.
Signature A variable-length field containing a digital
signature. The signature value is computed with
the hash algorithm and the signature algorithm,
as described in HA-id and SA-id. The signature
constructed by using the sender's private key
protects the following sequence of octets:
1. The DHCPv6 message header.
2. All DHCPv6 options including the Signature
option (fill the signature field with zeroes)
except for the Authentication Option.
The signature field MUST be padded, with all 0, to
the next octet boundary if its size is not a
multiple of 8 bits. The padding length depends on
the signature algorithm, which is indicated in the
SA-id field.
]]></artwork>
</figure>Note: if both signature and authentication option are
present, signature option does not protect the Authentication Option.
It allows the Authentication Option be created after signature has
been calculated and filled with the valid signature. It is because
both options need to apply hash algorithm to whole message, so there
must be a clear order and there could be only one last-created option.
In order to avoid update <xref target="RFC3315"></xref> because of
changing auth option, the authors chose not include authentication
option in the signature.</t>
</section>
<section anchor="TimeStampOption" title="Timestamp Option">
<t>The Timestamp option carries the current time on the sender. It
adds the anti-replay protection to the DHCPv6 messages. It is
optional.</t>
<t><figure>
<artwork><![CDATA[ 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OPTION_TIMESTAMP | option-len |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
| Timestamp (64-bit) |
| |
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
option-code OPTION_TIMESTAMP (TBA4).
option-len 8, in octets.
Timestamp The current time of day (NTP-format timestamp
[RFC5905] in UTC (Coordinated Universal Time), a
64-bit unsigned fixed-point number, in seconds
relative to 0h on 1 January 1900.). It can reduce
the danger of replay attacks.]]></artwork>
</figure></t>
</section>
<section anchor="StatusCodes" title="Status Codes">
<t>The following new status codes, see Section 5.4 of <xref
target="RFC3315"></xref> are defined. <list style="symbols">
<t>AlgorithmNotSupported (TBD5): indicates that the DHCPv6 server
does not support algorithms that sender used.</t>
<t>AuthenticationFail (TBD6): indicates that the DHCPv6 client
fails authentication check.</t>
<t>TimestampFail (TBD7): indicates the message from DHCPv6 client
fails the timestamp check.</t>
<t>SignatureFail (TBD8): indicates the message from DHCPv6 client
fails the signature check.</t>
</list></t>
</section>
</section>
<section title="Processing Rules and Behaviors">
<t>This section only covers the scenario where both DHCPv6 client and
DHCPv6 server are secure enabled.</t>
<section title="Processing Rules of Sender">
<t>The sender of a Secure DHCPv6 message could be a DHCPv6 server or a
DHCPv6 client.</t>
<t>The sender must have a public/private key pair in order to create
Secure DHCPv6 messages. The sender may also have a public key
certificate, which is signed by a CA assumed to be trusted by the
recipient, and its corresponding private key.</t>
<t>To support Secure DHCPv6, the Secure DHCPv6 enabled sender MUST
construct the DHCPv6 message following the rules defined in <xref
target="RFC3315"></xref>.</t>
<t>A Secure DHCPv6 message sent by a DHCPv6 server or a client, except
for Relay-reply messages, MUST either contain a Public Key option,
which MUST be constructed as explained in <xref
target="PKOption"></xref>, or a Certificate option, which MUST be
constructed as explained in <xref target="CertOption"></xref>.</t>
<t>A Secure DHCPv6 message, except for Relay-forward and Relay-reply
messages, MUST contain one and only one Signature option, which MUST
be constructed as explained in <xref target="SigOption"></xref>. It
protects the message header and all DHCPv6 options except for the
Authentication Option.</t>
<t>A Secure DHCPv6 message, except for Relay-forward and Relay-reply
messages, SHOULD contain one and only one Timestamp option, which MUST
be constructed as explained in <xref target="TimeStampOption"></xref>.
The Timestamp field SHOULD be set to the current time, according to
sender's real time clock.</t>
<t>A Relay-forward and relay-reply message MUST NOT contain any
additional Public Key or Certificate option or Signature Option or
Timestamp Option, aside from those present in the innermost
encapsulated messages from the client or server.</t>
<t>If the sender is a DHCPv6 client, in the failure cases, it receives
a Reply message with an error status code. The error status code
indicates the failure reason on the server side. According to the
received status code, the client MAY take follow-up action:</t>
<t><list style="symbols">
<t>Upon receiving an AlgorithmNotSupported error status code, the
client SHOULD resend the message protected with one of the
mandatory algorithms.</t>
<t>Upon receiving an AuthenticationFail error status code, the
client is not able to build up the secure communication with the
recipient. The client MAY switch to other public key certificate
if it has another one. But it SHOULD NOT retry with the same
certificate. However, if the client decides to retransmit using
the same certificate after receiving AuthenticationFail, it MUST
NOT retransmit immediately and MUST follow normal retransmission
routines defined in <xref target="RFC3315"></xref>.</t>
<t>Upon receiving a TimestampFail error status code, the client
MAY fall back to unsecured mode, or resend the message without a
Timestamp option. However, the DHCPv6 server MAY not accept the
message without a Timestamp option.</t>
<t>Upon receiving a SignatureFail error status code, the client
MAY resend the message following normal retransmission routines
defined in <xref target="RFC3315"></xref>.</t>
</list></t>
</section>
<section anchor="RuleRecipient" title="Processing Rules of Recipient ">
<t>The recipient of a Secure DHCPv6 message could be a DHCPv6 server
or a DHCPv6 client. In the failure cases, either DHCPv6 server or
client SHOULD NOT process received message, and the server SHOULD
reply a correspondent error status code, while the client does
nothing. The specific behavior depends on the configured local
policy.</t>
<t>When receiving a DHCPv6 message, except for Relay-Forward and
Relay-Reply messages, a Secure DHCPv6 enabled recipient SHOULD discard
any DHCPv6 messages that meet any of the following conditions:<list
style="symbols">
<t>the Signature option is absent,</t>
<t>multiple Signature options are present,</t>
<t>both the Public Key option and the Certificate option are
absent,</t>
<t>both the Public Key option and the Certificate option are
present.</t>
</list></t>
<t>In such failure, if the recipient is a DHCPv6 server, the server
SHOULD reply an UnspecFail (value 1, <xref target="RFC3315"></xref>)
error status code. If none of the Signature, Public Key or Certificate
options is present, the sender MAY be a legacy node or in unsecured
mode, then, the recipient MAY fall back to the unsecured DHCPv6 mode
if its local policy allows.</t>
<t>The recipient SHOULD first check the support of algorithms that
sender used. If not pass, the message is dropped. In such failure, if
the recipient is a DHCPv6 server, the server SHOULD reply an
AlgorithmNotSupported error status code, defined in <xref
target="StatusCodes"></xref>, back to the client. If both algorithms
are supported, the recipient then checks the authority of this sender.
The recipient SHOULD also use the same algorithms in the return
messages.</t>
<t>If a Certificate option is provided, the recipient SHOULD validate
the certificate according to the rules defined in <xref
target="RFC5280"></xref>. An implementation may create a local trust
certificate record for verified certificates in order to avoid
repeated verification procedure in the future. A certificate that
finds a match in the local trust certificate list is treated as
verified.</t>
<t>If a Public Key option is provided, the recipient SHOULD validate
it by finding a matching public key from the local trust public key
list, which is pre-configured or recorded from previous communications
(TOFU). A local trust public key list is a data table maintained by
the recipient. It stores public keys from all trustworthy senders.</t>
<t>The message that fails authentication check MUST be dropped. In
such failure, the DHCPv6 server SHOULD reply an AuthenticationFail
error status code, defined in <xref target="StatusCodes"></xref>, back
to the client.</t>
<t>The recipient MAY choose to further process messages from a sender
when there is no matched public key. By recording the public key, when
the first time it is seen, the recipient can make a Trust On First Use
that the sender is trustworthy. The circumstances under which this
might be done are out of scope for this document. </t>
<t>At this point, the recipient has either recognized the
authentication of the sender, or decided to drop the message. The
recipient MUST now authenticate the sender by verifying the signature
and checking timestamp (see details in <xref
target="timestampCheck"></xref>), if there is a Timestamp option. The
order of two procedures is left as an implementation decision. It is
RECOMMENDED to check timestamp first, because signature verification
is much more computationally expensive. Depending on server's local
policy, the message without a Timestamp option MAY be acceptable or
rejected. If the server rejects such a message, a TimestampFail error
status code, defined in <xref target="StatusCodes"></xref>, should be
sent back to the client. The reply message that carries the
TimestampFail error status code SHOULD NOT carry a timestamp
option.</t>
<t>The signature field verification MUST show that the signature has
been calculated as specified in <xref target="SigOption"></xref>. Only
the messages that get through both the signature verifications and
timestamp check (if there is a Timestamp option) are accepted as
secured DHCPv6 messages and continue to be handled for their contained
DHCPv6 options as defined in <xref target="RFC3315"></xref>. Messages
that do not pass the above tests MUST be discarded or treated as
unsecured messages. In the case the recipient is DHCPv6 server, the
DHCPv6 server SHOULD reply a SignatureFail error status code, defined
in <xref target="StatusCodes"></xref>, for the signature verification
failure; or a TimestampFail error status code, defined in <xref
target="StatusCodes"></xref>, for the timestamp check failure, back to
the client.</t>
<t>Furthermore, the node that supports the verification of the Secure
DHCPv6 messages MAY impose additional constraints for the
verification. For example, it may impose limits on minimum and maximum
key lengths.</t>
<t><list style="hanging">
<t hangText="Minbits">The minimum acceptable key length for public
keys. An upper limit MAY also be set for the amount of computation
needed when verifying packets that use these security
associations. The appropriate lengths SHOULD be set according to
the signature algorithm and also following prudent cryptographic
practice. For example, minimum length 1024 and upper limit 2048
may be used for RSA <xref target="RSA"></xref>.</t>
</list>A Relay-forward or Relay-reply message with any Public Key,
Certificate or the Signature option is invalid. The message MUST be
discarded silently.</t>
</section>
<section title="Processing Rules of Relay Agent">
<t>To support Secure DHCPv6, relay agents just need to follow the same
processing rules defined in <xref target="RFC3315"></xref>. There is
nothing more the relay agents have to do, either verify the messages
from client or server, or add any secure DHCPv6 options. Actually, by
definition in this document, relay agents SHOULD NOT add any secure
DHCPv6 options.</t>
</section>
<section anchor="timestampCheck" title="Timestamp Check">
<t>In order to check the Timestamp option, defined in <xref
target="TimeStampOption"></xref>, recipients SHOULD be configured with
an allowed timestamp Delta value, a "fuzz factor" for comparisons, and
an allowed clock drift parameter. The recommended default value for
the allowed Delta is 300 seconds (5 minutes); for fuzz factor 1
second; and for clock drift, 0.01 second.</t>
<t>Note: the Timestamp mechanism is based on the assumption that
communication peers have roughly synchronized clocks, with certain
allowed clock drift. So, accurate clock is not necessary. If one has a
clock too far from the current time, the timestamp mechanism would not
work.</t>
<t>To facilitate timestamp checking, each recipient SHOULD store the
following information for each sender, from which at least one
accepted secure DHCPv6 message is successfully verified (for both
timestamp check and signature verification):</t>
<t><list style="symbols">
<t>The receive time of the last received and accepted DHCPv6
message. This is called RDlast.</t>
<t>The timestamp in the last received and accepted DHCPv6 message.
This is called TSlast.</t>
</list>A verified (for both timestamp check and signature
verification) secure DHCPv6 message initiates the update of the above
variables in the recipient's record.</t>
<t>Recipients MUST check the Timestamp field as follows:</t>
<t><list style="symbols">
<t>When a message is received from a new peer (i.e., one that is
not stored in the cache), the received timestamp, TSnew, is
checked, and the message is accepted if the timestamp is recent
enough to the reception time of the packet, RDnew:<list
style="empty">
<t>-Delta < (RDnew - TSnew) < +Delta</t>
</list><vspace blankLines="1" />After the signature verification
also succeeds, the RDnew and TSnew values SHOULD be stored in the
cache as RDlast and TSlast.</t>
<t>When a message is received from a known peer (i.e., one that
already has an entry in the cache), the timestamp is checked
against the previously received Secure DHCPv6 message:<list
style="empty">
<t>TSnew + fuzz > TSlast + (RDnew - RDlast) x (1 - drift) -
fuzz</t>
</list><vspace blankLines="1" />If this inequality does not hold
or RDnew < RDlast, the recipient SHOULD silently discard the
message. If, on the other hand, the inequality holds, the
recipient SHOULD process the message. <vspace
blankLines="1" />Moreover, if the above inequality holds and TSnew
> TSlast, the recipient SHOULD update RDlast and TSlast after
the signature verification also successes. Otherwise, the
recipient MUST NOT update RDlast or TSlast.</t>
</list>An implementation MAY use some mechanism such as a timestamp
cache to strengthen resistance to replay attacks. When there is a very
large number of nodes on the same link, or when a cache filling attack
is in progress, it is possible that the cache holding the most recent
timestamp per sender will become full. In this case, the node MUST
remove some entries from the cache or refuse some new requested
entries. The specific policy as to which entries are preferred over
others is left as an implementation decision.</t>
<t>An implementation MAY statefully record the latest timestamps from
senders. In such implementation, the timestamps MUST be strictly
monotonously increasing. This is reasonable given that DHCPv6 messages
are rarely misordered.</t>
</section>
</section>
<section anchor="DeployConsider" title="Deployment Consideration">
<t>This document defines two directions of authentication:
authentication based on client's public key certificate and
authentication based on leap of faith to server's public key.</t>
<section title="Authentication on a client">
<t>For clients, DHCPv6 authentication generally means verifying
whether the sender of DHCPv6 messages is a legal DHCPv6 server and
verifying whether the message has been modified during transmission.
Because the client may have to validate the authentication in the
condition of without connectivity wider than link-local,
authentication with certificates may not always be feasible. So, this
document only sticks on Leaf of Faith mode, to make sure the client
talks to the same previous server.</t>
<t>Message integrity is provided. But there is a chance for the client
to incorrectly trust a malicious server at the beginning of the first
session with the server (and therefore keep trusting it thereafter).
But the leap of faith mechanim guarantees the subsequent messages are
sent by the same previous server, and therefore narrows the attack
scope. This may make sense if the network can be reasonably considered
secure and requesting pre-configuration is deemed to be infeasible. A
small home network would be an example of such cases.</t>
<t>For environments that are neither controlled nor really
trustworthy, such as a network in a cafeteria, while the leap of faith
mode, i.e., silently trusting the server at the first time, would be
too insecure. But some middle ground might be justified, such as
requiring human intervention at the point of the leap of faith.</t>
</section>
<section title="Authentication on a server">
<t>As for authentication on a server, there are several different
scenarios to consider, each of which has different applicability
issues. If the server allows the leap of faith mode, any malicious
user can pretend to be a new legitimate client. While the server can
always be considered to have connectivity to validate certificate, it
is feasible to check client certificates.</t>
<t>Network administrators may wish to constrain the allocation of
addresses to authorized hosts to avoid denial of service attacks in
"hostile" environments where the network medium is not physically
secured, such as wireless networks or college residence halls. A
server may have to selectively serve a specific client or deny
specific clients depending on the identity of the client in a
controlled environment, like a corporate intranet. But the support
from skilled staff or administrator may be required to set up the
clients.</t>
</section>
</section>
<section anchor="Security" title="Security Considerations">
<t>This document provides new security features to the DHCPv6
protocol.</t>
<t>Using public key based security mechanism and its verification
mechanism in DHCPv6 message exchanging provides the authentication and
data integrity protection. Timestamp mechanism provides anti-replay
function.</t>
<t>The Secure DHCPv6 mechanism is based on the pre-condition that the
recipient knows the public key of the sender or the sender's public key
certificate can be verified through a trust CA. Clients may discard the
DHCPv6 messages from unknown/unverified servers, which may be fake
servers; or may prefer DHCPv6 messages from known/verified servers over
unsigned messages or messages from unknown/unverified servers. The
pre-configuration operation also needs to be protected, which is out of
scope. The deployment of PKI is also out of scope.</t>
<t>When a recipient first encounters a new public key, it may also store
the key using a Trust On First Use policy. If the sender that used that
public key is in fact legitimate, then all future communication with
that sender can be protected by storing the public key. This does not
provide complete security, but it limits the opportunity to mount an
attack on a specific recipient to the first time it communicates with a
new sender.</t>
<t>Downgrade attacks cannot be avoided if nodes are configured to accept
both secured and unsecured messages. A future specification may provide
a mechanism on how to treat unsecured DHCPv6 messages.</t>
<t><xref target="RFC6273"></xref> has analyzed possible threats to the
hash algorithms used in SEND. Since the Secure DHCPv6 defined in this
document uses the same hash algorithms in similar way to SEND, analysis
results could be applied as well: current attacks on hash functions do
not constitute any practical threat to the digital signatures used in
the signature algorithm in the Secure DHCPv6.</t>
<t>A server, whose local policy accepts messages without a Timestamp
option, may have to face the risk of replay attacks.</t>
<t>A window of vulnerability for replay attacks exists until the
timestamp expires. Secure DHCPv6 nodes are protected against replay
attacks as long as they cache the state created by the message
containing the timestamp. The cached state allows the node to protect
itself against replayed messages. However, once the node flushes the
state for whatever reason, an attacker can re-create the state by
replaying an old message while the timestamp is still valid. In
addition, the effectiveness of timestamps is largely dependent upon the
accuracy of synchronization between communicating nodes. However, how
the two communicating nodes can be synchronized is out of scope of this
work.</t>
<t>Attacks against time synchronization protocols such as NTP [RFC5905]
may cause Secure DHCPv6 nodes to have an incorrect timestamp value. This
can be used to launch replay attacks, even outside the normal window of
vulnerability. To protect against these attacks, it is recommended that
Secure DHCPv6 nodes keep independently maintained clocks or apply
suitable security measures for the time synchronization protocols.</t>
<t>One more consideration is that this protocol does reveal additional
client information in their certificate. It means less privacy. In
current practice, the client privacy and the client authentication are
mutually exclusive.</t>
</section>
<section anchor="IANA" title="IANA Considerations">
<t>This document defines four new DHCPv6 <xref target="RFC3315"></xref>
options. The IANA is requested to assign values for these four options
from the DHCPv6 Option Codes table of the DHCPv6 Parameters registry
maintained in http://www.iana.org/assignments/dhcpv6-parameters. The
four options are:</t>
<t><list style="empty">
<t>The Public Key Option (TBA1), described in <xref
target="PKOption"></xref>.</t>
<t>The Certificate Option (TBA2), described in <xref
target="CertOption"></xref>.</t>
<t>The Signature Option (TBA3), described in <xref
target="SigOption"></xref>.</t>
<t>The Timestamp Option (TBA4),described in <xref
target="TimeStampOption"></xref>.</t>
</list>The IANA is also requested to add two new registry tables to
the DHCPv6 Parameters registry maintained in
http://www.iana.org/assignments/dhcpv6-parameters. The two tables are
the Hash Algorithm for Secure DHCPv6 table and the Signature Algorithm
for Secure DHCPv6 table.</t>
<t>Initial values for these registries are given below. Future
assignments are to be made through Standards Action <xref
target="RFC5226"></xref>. Assignments for each registry consist of a
name, a value and a RFC number where the registry is defined.</t>
<t>Hash Algorithm for Secure DHCPv6. The values in this table are 8-bit
unsigned integers. The following initial values are assigned for Hash
Algorithm for Secure DHCPv6 in this document:</t>
<t><figure>
<artwork><![CDATA[ Name | Value | RFCs
-------------------+---------+--------------
SHA-256 | 0x01 | this document
SHA-512 | 0x02 | this document
]]></artwork>
</figure>Signature Algorithm for Secure DHCPv6. The values in this
table are 8-bit unsigned integers. The following initial values are
assigned for Signature Algorithm for Secure DHCPv6 in this document:</t>
<t><figure>
<artwork><![CDATA[ Name | Value | RFCs
-------------------+---------+--------------
RSASSA-PKCS1-v1_5 | 0x01 | this document
]]></artwork>
</figure>IANA is requested to assign the following new DHCPv6 Status
Codes, defined in <xref target="StatusCodes"></xref>, in the DHCPv6
Parameters registry maintained in
http://www.iana.org/assignments/dhcpv6-parameters:</t>
<t><figure>
<artwork><![CDATA[ Code | Name | Reference
---------+-----------------------+--------------
TBD5 | AlgorithmNotSupported | this document
TBD6 | AuthenticationFail | this document
TBD7 | TimestampFail | this document
TBD8 | SignatureFail | this document
]]></artwork>
</figure></t>
</section>
<section anchor="Acknowledgments" title="Acknowledgements">
<t>The authors would like to thank Bernie Volz, Ted Lemon, Ralph Droms,
Jari Arkko, Sean Turner, Stephen Kent, Thomas Huth, David Schumacher,
Francis Dupont, Tomek Mrugalski, Gang Chen, Qi Sun, Suresh Krishnan,
Fred Templin and other members of the IETF DHC working group for their
valuable comments.</t>
<t>This document was produced using the xml2rfc tool <xref
target="RFC2629"></xref>.</t>
</section>
<section anchor="changes" title="Change log [RFC Editor: Please remove]">
<t>draft-ietf-dhc-sedhcpv6-06: remove the limitation that only clients
use PKI- certificates and only servers use public keys. The new text
would allow clients use public keys and servers use PKI-certificates</t>
<t>draft-ietf-dhc-sedhcpv6-05: addressed comments from mail list that
responsed to the second WGLC.</t>
<t>draft-ietf-dhc-sedhcpv6-04: addressed comments from mail list. Making
timestamp an independent and optional option. Reduce the serverside
authentication to base on only client's certificate. Reduce the
clientside authentication to only Leaf of Faith base on server's public
key. 2014-09-26.</t>
<t>draft-ietf-dhc-sedhcpv6-03: addressed comments from WGLC. Added a new
section "Deployment Consideration". Corrected the Public Key Field in
the Public Key Option. Added consideration for large DHCPv6 message
transmission. Added TimestampFail error code. Refined the retransmission
rules on clients. 2014-06-18.</t>
<t>draft-ietf-dhc-sedhcpv6-02: addressed comments (applicability
statement, redesign the error codes and their logic) from IETF89 DHC WG
meeting and volunteer reviewers. 2014-04-14.</t>
<t>draft-ietf-dhc-sedhcpv6-01: addressed comments from IETF88 DHC WG
meeting. Moved Dacheng Zhang from acknowledgement to be co-author.
2014-02-14.</t>
<t>draft-ietf-dhc-sedhcpv6-00: adopted by DHC WG. 2013-11-19.</t>
<t>draft-jiang-dhc-sedhcpv6-02: removed protection between relay agent
and server due to complexity, following the comments from Ted Lemon,
Bernie Volz. 2013-10-16.</t>
<t>draft-jiang-dhc-sedhcpv6-01: update according to review comments from
Ted Lemon, Bernie Volz, Ralph Droms. Separated Public Key/Certificate
option into two options. Refined many detailed processes.
2013-10-08.</t>
<t>draft-jiang-dhc-sedhcpv6-00: original version, this draft is a
replacement of draft-ietf-dhc-secure-dhcpv6, which reached IESG and dead
because of consideration regarding to CGA. The authors followed the
suggestion from IESG making a general public key based mechanism.
2013-06-29.</t>
</section>
</middle>
<back>
<references title="Normative References">
<?rfc include='reference.RFC.2119'?>
<?rfc include='reference.RFC.2460'?>
<?rfc include='reference.RFC.3279'?>
<?rfc include='reference.RFC.3315'?>
<?rfc include='reference.RFC.4055'?>
<?rfc include='reference.RFC.4443'?>
<?rfc include='reference.RFC.4491'?>
<?rfc include='reference.RFC.5280'?>
<?rfc include='reference.RFC.5905'?>
<?rfc include='reference.RFC.7296'?>
</references>
<references title="Informative References">
<reference anchor="RSA">
<front>
<title>RSA Encryption Standard, Version 2.1, PKCS 1</title>
<author fullname="">
<organization>RSA Laboratories</organization>
</author>
<date month="November" year="2002" />
</front>
</reference>
<?rfc include='reference.RFC.2629'?>
<?rfc include='reference.RFC.4270'?>
<?rfc include='reference.RFC.5226'?>
<?rfc include='reference.RFC.6273'?>
</references>
</back>
</rfc>
| PAFTECH AB 2003-2026 | 2026-04-24 01:52:40 |