One document matched: draft-ietf-anima-bootstrapping-keyinfra-04.xml
<?xml version="1.0" encoding="US-ASCII"?>
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
<!ENTITY I-D.behringer-autonomic-network-framework PUBLIC "" "https://xml2rfc.ietf.org/public/rfc/bibxml3/reference.I-D.behringer-homenet-trust-bootstrap.xml">
<!ENTITY I-D.draft-ietf-anima-autonomic-control-plane PUBLIC "" "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.behringer-anima-autonomic-control-plane.xml">
<!ENTITY RFC2119 SYSTEM "https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2119.xml">
<!ENTITY RFC7030 SYSTEM "https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7030.xml">
<!ENTITY RFC5280 SYSTEM "https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5280.xml">
<!ENTITY RFC5652 SYSTEM "https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5652.xml">
<!ENTITY RFC7575 SYSTEM "https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7575.xml">
]>
<rfc category="info" docName="draft-ietf-anima-bootstrapping-keyinfra-04"
ipr="trust200902">
<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>
<?rfc toc="yes" ?>
<?rfc compact="yes" ?>
<?rfc symrefs="yes" ?>
<?rfc sortrefs="yes"?>
<?rfc iprnotified="no" ?>
<?rfc strict="yes" ?>
<front>
<title abbrev="BRewSKI">Bootstrapping Remote Secure Key Infrastructures
(BRSKI)</title>
<author fullname="Max Pritikin" initials="M." surname="Pritikin">
<organization>Cisco</organization>
<address>
<email>pritikin@cisco.com</email>
</address>
</author>
<author fullname="Michael C. Richardson" initials="M."
surname="Richardson">
<organization abbrev="SSW">Sandelman Software Works</organization>
<address>
<email>mcr+ietf@sandelman.ca</email>
<uri>http://www.sandelman.ca/</uri>
</address>
</author>
<author fullname="Michael H. Behringer" initials="M.H."
surname="Behringer">
<organization>Cisco</organization>
<address>
<email>mbehring@cisco.com</email>
</address>
</author>
<author fullname="Steinthor Bjarnason" initials="S." surname="Bjarnason">
<organization>Cisco</organization>
<address>
<email>sbjarnas@cisco.com</email>
</address>
</author>
<author initials="K.W." surname="Watsen" fullname="Kent Watsen">
<organization>Juniper Networks</organization>
<address>
<email>kwatsen@juniper.net</email>
</address>
</author>
<date year="2016" />
<area>Operations and Management</area>
<workgroup>ANIMA WG</workgroup>
<abstract>
<t>This document specifies automated bootstrapping of a remote secure
key infrastructure (BRSKI) using vendor installed X.509 certificate, in
combination with a vendor authorized service on the Internet.
Bootstrapping a new device can occur using a routable address and a
cloud service, or using only link-local connectivity, or on
limited/disconnected networks. Support for lower security models,
including devices with minimal identity, is described for legacy reasons
but not encouraged. Bootstrapping is complete when the cryptographic
identity of the new key infrastructure is successfully deployed to the
device but the established secure connection can be used to deploy a
locally issued certificate to the device as well.</t>
</abstract>
</front>
<middle>
<section title="Introduction">
<t>To literally "pull yourself up by the bootstraps" is an impossible
action. Similarly the secure establishment of a key infrastructure
without external help is also an impossibility. Today it is accepted
that the initial connections between nodes are insecure, until key
distribution is complete, or that domain-specific keying material is
pre-provisioned on each new device in a costly and non-scalable manner.
This document describes a zero-touch approach to bootstrapping an entity
by securing the initial distribution of key material using third-party
issued X.509 certificates and cryptographically signed "vouchers" issued
by a new form of cloud service.</t>
<t>The two sides of an association being bootstrapped authenticate each
other and then determine appropriate authorization. This process is
described as four distinct steps between the existing domain and the
device, or "pledge", being added:</t>
<t><list style="symbols">
<t>Pledge authentication: "Who is this? What is its identity?"</t>
<t>Pledge authorization: "Is it mine? Do I want it? What are the
chances it has been compromised?"</t>
<t>Domain authentication: "What is this domain's claimed
identity?"</t>
<t>Domain authorization: "Should I join it?"</t>
</list></t>
<t>A precise answer to these questions can not be obtained without
leveraging an established key infrastructure(s). The pledge's decisions
are made according to verified communication with a trusted third-party.
The domain's decisions are made by comparing the pledge's authenticated
identity against domain information such as a configured list of
purchased devices supplimented by information provided by a trusted
third-party. The third-party is not required to provide sales channel
ownership tracking nor is it required to authenticate the domain.</t>
<t>Optimal security is achieved with X.509 certificates on each Pledge,
accompanied by a third-party (e.g., vendor, manufacturer or integrator)
Internet based service for verification. Bootstrapping concepts run to
completion with less requirements, but are then less secure. A domain
can choose to accept lower levels of security when a trusted third-party
is not available so that bootstrapping proceeds even at the risk of
reduced security. Only the domain can make these decisions based on
administrative input and known behavior of the pledge.</t>
<t>The result of bootstrapping is that a domain specific key
infrastructure is deployed. Since X.509 PKI certificates are used for
identifying the pledge, and the public key of the domain identity is
leveraged during communications with an Internet based service, which is
itself authenticated using HTTPS, bootstrapping of a domain specific
Public Key Infrastructure (PKI) is described. Sufficient agility to
support bootstrapping alternative key infrastructures (such as symmetric
key solutions) is considered although no such alternate key
infrastructure is described.</t>
<section title="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>.</t>
<t>The following terms are defined for clarity:</t>
<t><list style="hanging">
<t hangText="DomainID:">The domain identity is the 160-bit SHA-1
hash of the BIT STRING of the subjectPublicKey of the domain trust
anchor that is stored by the Domain CA. This is consistent with
the Certification Authority subject key identifier (<xref
target="RFC5280">Section 4.2.1.2</xref>) of the Domain CA's self
signed root certificate. (A string value bound to the Domain CA's
self signed root certificate subject and issuer fields is often
colloquially used as a humanized identity value but during
protocol discussions the more exact term as defined here is
used).</t>
<t hangText="drop ship:">The physical distribution of equipment
containing the "factory default" configuration to a final
destination. In zero-touch scenarios there is no staging or
pre-configuration during drop-ship.</t>
<t hangText="imprint:">The process where a device obtains the
cryptographic key material to identify and trust future
interactions with a network. This term is taken from Konrad
Lorenz's work in biology with new ducklings: during a critical
period, the duckling would assume that anything that looks like a
mother duck is in fact their mother. An equivalent for a device is
to obtain the fingerprint of the network's root certification
authority certificate. A device that imprints on an attacker
suffers a similar fate to a duckling that imprints on a hungry
wolf. Securely imprinting is a primary focus of this
document.<xref target="imprinting"></xref>. The analogy to
Lorenz's work was first noted in <xref
target="Stajano99theresurrecting"></xref>.</t>
<t hangText="enrollment:">The process where a device presents key
material to a network and acquires a network specific identity.
For example when a certificate signing request is presented to a
certification authority and a certificate is obtained in
response.</t>
<t hangText="Pledge:">The prospective device, which has an
identity installed by a third-party (e.g., vendor, manufacturer or
integrator).</t>
<t hangText="Voucher">A signed statement from the MASA service
that indicates to a Pledge the cryptographic identity of the
Registrar it should trust. There are different types of vouchers
depending on how that trust verified.</t>
<t hangText="Audit Voucher:">A voucher from the MASA service that
indicates that the bootstrapping event has been successfully
logged. The Registrar is primarily responsible for verifying the
logs and ensuring domain network security.</t>
<t hangText="Ownership Voucher:">A voucher from the MASA service
that indicates the explicit owner identity. The MASA is primarily
responsible for tracking ownership using out-of-band sales channel
integration (the definition of which is out-of-scope of this
document). It is defined in <xref
target="I-D.ietf-netconf-zerotouch"></xref>.</t>
<t hangText="Domain:">The set of entities that trust a common key
infrastructure trust anchor. This includes the Proxy, Registrar,
Domain Certificate Authority, Management components and any
existing entity that is already a member of the domain.</t>
<t hangText="Domain CA:">The domain Certification Authority (CA)
provides certification functionalities to the domain. At a minimum
it provides certification functionalities to a Registrar and
stores the trust anchor that defines the domain. Optionally, it
certifies all elements.</t>
<t hangText="Registrar:">A representative of the domain that is
configured, perhaps autonomically, to decide whether a new device
is allowed to join the domain. The administrator of the domain
interfaces with a Registrar to control this process. Typically a
Registrar is "inside" its domain.</t>
<t hangText="Proxy:">A domain entity that helps the pledge join
the domain. A Proxy facilitates communication for devices that
find themselves in an environment where they are not provided
connectivity until after they are validated as members of the
domain. The pledge is unaware that they are communicating with a
proxy rather than directly with a Registrar.</t>
<t hangText="MASA Service:">A third-party Manufacturer Authorized
Signing Authority (MASA) service on the global Internet. The MASA
provides a repository for audit log information concerning privacy
protected bootstrapping events. It does not track ownership.</t>
<t hangText="Ownership Tracker">An Ownership Tracker service on
the global internet. The Ownership Tracker uses business processes
to accurately track ownership of all devices shipped against
domains that have purchased them. Although optional this component
allows vendors to provide additional value in cases where their
sales and distribution channels allow for accurately tracking of
such ownership.</t>
<t hangText="IDevID">An Initial Device Identity X.509 certificate
installed by the vendor on new equipment. The <xref
target="IDevID"></xref> certificate format is the primary example.
In particular the X.509 certificate needs to contain the device's
serial number in a well known location in order to perform white
list operations and in order to extract it for inclusion in
messages to the MASA service. The subject field’s DN
encoding MUST include the “serialNumber” attribute
with the device’s unique serial number.</t>
</list></t>
</section>
<section title="Scope of solution">
<t>Questions have been posed as to whether this solution is suitable
in general for Internet of Things (IoT) networks. This depends on the
capabilities of the devices in question. The terminology of <xref
target="RFC7228"></xref> is best used to describe the boundaries.</t>
<t>The entire solution described in this document is aimed in general
at non-constrained (i.e. class 2+) devices operating on a
non-Challenged network. The entire solution described here is not
intended to be useable as-is by constrained devices operating on
challenged networks (such as 802.15.4 LLNs).</t>
<t>In many target applications, the systems involved are large router
platforms with multi-gigabit inter-connections, mounted in controlled
access data centers. But this solution is not exclusive to the large,
it is intended to scale to thousands of devices located in hostile
environments, such as ISP provided CPE devices which are drop-shipped
to the end user. The situation where an order is fulfilled from
distributed warehouse from a common stock and shipped directly to the
target location at the request of the domain owner is explicitly
supported. That stock ("SKU") could be provided to a number of
potential domain owners, and the eventual domain owner will not know
a-priori which device will go to which location.</t>
<t>The bootstraping process can take minutes to complete depending on
the network infrastructure and device processing speed. The network
communication itself is not optimized for speed; the discovery process
allows for the Pledge to avoid broadcasting for privacy reasons. This
protocol is not intended for low latency handoffs.</t>
<t>Specifically, there are protocol aspects described here which might
result in congestion collapse or energy-exhaustion of intermediate
battery powered routers in an LLN. Those types of networks SHOULD NOT
use this solution. These limitations are predominately related to the
large credential and key sizes required for device authentication.
Defining symmetric key techniques that meet the operational
requirements is out-of-scope but the underlying protocol operations
(TLS handshake and signing structures) have sufficient algorithm
agility to support such techniques when defined.</t>
<t>The imprint protocol described here could, however, be used by
non-energy constrained devices joining a non-constrained network (for
instance, smart light bulbs are usually mains powered, and speak
802.11). It could also be used by non-constrained devices across a
non-energy constrained, but challenged network (such as 802.15.4).</t>
<t>The use of an IDevID that is consistant with <xref
target="IDevID"></xref> allows for alignment with 802.1X network
access control methods which could need to complete before
bootstrapping can be initiated. This document presumes that network
access control has either already occured, is not required, or is
integrated by the proxy and registrar in such a way that the device
itself does not need to be aware of the details. Further integration
is not in scope.</t>
<t>Some aspects are in scope for constrained devices on challenged
networks: the certificate contents, and the process by which the four
questions above are resolved is in scope. It is simply the actual
on-the-wire imprint protocol which is likely inappropriate.</t>
</section>
<section title="Trust bootstrap">
<t>The imprint protocol results in a secure relationship between a
domain Registrar and the Pledge. If the new device is sufficiently
constrained that the ACE protocol should be leveraged for operation,
(see <xref target="I-D.ietf-ace-actors"></xref>), and the domain
registrar is also the Client Authorization Server or the Authorization
Server, then it may be appropriate to use this secure channel to
exchange ACE tokens.</t>
</section>
</section>
<section title="Architectural Overview">
<t>The logical elements of the bootstrapping framework are described in
this section. Figure 1 provides a simplified overview of the components.
Each component is logical and may be combined with other components as
necessary.</t>
<t></t>
<figure>
<artwork><![CDATA[ .
.+------------------------+
+--------------Drop Ship-------------->.| Vendor Service |
| .+------------------------+
| .| M anufacturer| |
| .| A uthorized |Ownership|
| .| S igning |Tracker |
| .| A uthority | |
| .+--------------+---------+
| .............. ^
V |
+-------+ ............................................|...
| | . | .
| | . +------------+ +-----------+ | .
| | . | | | | | .
| | . | | | <-------+ .
| | . | Proxy | | Registrar | .
| <--------> <-------> | .
| New | . | | | | .
| Entity| . +------------+ +-----+-----+ .
| | . | .
| | . +-----------------+----------+ .
| | . | Domain Certification | .
| | . | Authority | .
+-------+ . | Management and etc | .
. +----------------------------+ .
. .
................................................
"Domain" components
]]></artwork>
<postamble>Figure 1</postamble>
</figure>
<t>We assume a multi-vendor network. In such an environment there could
be a MASA or Ownership Tracker for each vendor that supports devices
following this document's specification, or an integrator could provide
a MASA service for all devices. It is unlikely that an integrator could
provide Ownership Tracking services for multiple vendors.</t>
<t>This document describes a secure zero-touch approach to bootstrapping
a key infrastructure; if certain devices in a network do not support
this approach, they can still be bootstrapped manually. Although manual
deployment is not scalable and is not a focus of this document the
necessary mechanisms are called out in this document to ensure such edge
conditions are covered by the architectural and protocol models.</t>
</section>
<section title="Functional Overview">
<t>Entities behave in an autonomic fashion. They discover each other and
autonomically bootstrap into a key infrastructure delineating the
autonomic domain. See <xref target="RFC7575"></xref> for more
information.</t>
<t>This section details the state machine and operational flow for each
of the main three entities. The pledge, the domain (primarily a
Registrar) and the MASA service.</t>
<t>A representative flow is shown in Figure 2:</t>
<figure>
<artwork><![CDATA[
+--------+ +---------+ +------------+ +------------+
| Pledge | | Circuit | | Domain | | Vendor |
| | | Proxy | | Registrar | | Service |
| | | | | | | (Internet |
+--------+ +---------+ +------------+ +------------+
| | | |
|<-RFC3927 IPv4 adr | | |
or|<-RFC4862 IPv6 adr | | |
| | | |
|-------------------->| | |
| optional: mDNS query| | |
| RFC6763/RFC6762 | | |
| | | |
|<--------------------| | |
| mDNS broadcast | | |
| response or periodic| | |
| | | |
|<------------------->C<----------------->| |
| TLS via the Circuit Proxy | |
|<--Registrar TLS server authentication---| |
[PROVISIONAL accept of server cert] | |
P---X.509 client authentication---------->| |
P | | |
P---Request Voucher (include nonce)------>| |
P | | |
P | /---> | |
P | | [accept device?] |
P | | [contact Vendor] |
P | | |--Pledge ID-------->|
P | | |--Domain ID-------->|
P | | |--optional:nonce--->|
P | | | [extract DomainID]
P | | | |
P | optional: | [update audit log]
P | |can | |
P | |occur | |
P | |in | |
P | |advance | |
P | | | |
P | | |<-device audit log--|
P | | |<- voucher ---------|
P | \----> | |
P | | |
P | [verify audit log and voucher] |
P | | |
P<------voucher---------------------------| |
[verify voucher ] | | |
[verify provisional cert ]| | |
| | | |
|---------------------------------------->| |
| Continue with RFC7030 enrollment | |
| using now bidirectionally authenticated | |
| TLS session. | | |
| | | |
| | | |
| | | |
]]></artwork>
<postamble>Figure 2</postamble>
</figure>
<t></t>
<section title="Behavior of a Pledge">
<t>A pledge that has not yet been bootstrapped attempts to find a
local domain and join it. A pledge MUST NOT automatically initiate
bootstrapping if it has already been configured or is in the process
of being configured.</t>
<t>States of a pledge are as follows:</t>
<t></t>
<figure>
<artwork><![CDATA[
+--------------+
| Start |
| |
+------+-------+
|
+------v-------+
| Discover |
+------------> |
| +------+-------+
| |
| +------v-------+
| | Identity |
^------------+ |
| rejected +------+-------+
| |
| +------v-------+
| | Request |
| | Join |
| +------+-------+
| |
| +------v-------+
| | Imprint | Optional
^------------+ <--+Manual input
| Bad Vendor +------+-------+
| response |
| +------v-------+
| | Enroll |
^------------+ |
| Enroll +------+-------+
| Failure |
| +------v-------+
| | Being |
^------------+ Managed |
Factory +--------------+
reset
]]></artwork>
<postamble>Figure 3</postamble>
</figure>
<t></t>
<t>State descriptions for the pledge are as follows:</t>
<t><list style="numbers">
<t>Discover a communication channel to a Registrar.</t>
<t>Identify itself. This is done by presenting an IDevID X.509
credential to the discovered Registrar (via the Proxy) in a TLS
handshake. (The Registrar credentials are only provisionally
accepted at this time).</t>
<t>Requests to Join the discovered Registrar. A unique nonce is
included ensuring that any responses can be associated with this
particular bootstrapping attempt.</t>
<t>Imprint on the Registrar. This requires verification of the
vendor service provided "Audit" or "Ownership" Voucher. Either of
these responses contains sufficient information for the pledge to
complete authentication of a Registrar. (The pledge can now finish
authentication of the Registrar TLS server certificate)</t>
<t>Enroll by accepting the domain specific information from a
Registrar, and by obtaining a domain certificate from a Registrar
using a standard enrollment protocol, e.g. Enrollment over Secure
Transport (EST) <xref target="RFC7030"></xref>.</t>
<t>The Pledge is now a member of, and can be managed by, the
domain and will only repeat the discovery aspects of bootstrapping
if it is returned to factory default settings.</t>
</list>The following sections describe each of these steps in more
detail.</t>
<section anchor="ProxyDiscovery" title="Discovery">
<t>The result of discovery is a logical communication with a
Registrar, through a Proxy. The Proxy is transparent to the Pledge
but is always assumed to exist.</t>
<t>To discover the Registrar the Pledge performs the following
actions:</t>
<t><list style="letters">
<t>MUST: Obtains a local address using either IPv4 or IPv6
methods as described in <xref target="RFC4862"></xref> IPv6
Stateless Address AutoConfiguration or <xref
target="RFC3927"></xref> Dynamic Configuration of IPv4
Link-Local Addresses. The Plege MAY obtain an IP address via
DHCP [RFC2131]. The DHCP provided parameters for the Domain Name
System can be used to perform step (d) DNS operations if all
local discovery attempts fail (see below).</t>
<t>MUST: Performs DNS-based Service Discovery <xref
target="RFC6763"></xref> over Multicast DNS <xref
target="RFC6762"></xref> searching for the service
"_bootstrapks._tcp.local.". To prevent unaccceptable levels of
network traffic the congestion avoidance mechanisms specified in
<xref target="RFC6762"></xref> section 7 MUST be followed. The
Pledge SHOULD listen for an unsolicited broadcast response as
described in <xref target="RFC6762"></xref>. This allows devices
to avoid announcing their presence via mDNS broadcasts and
instead silently join a network by watching for periodic
unsolicited broadcast responses.</t>
<t>MAY: Performs DNS-based Service Discovery [RFC6763] over
normal DNS operations. The service searched for is
"_bootstrapks._tcp.example.com". In this case the domain
"example.com" is discovered as described in <xref
target="RFC6763"></xref> section 11.</t>
<t>MAY: If no local bootstrapks service is located using the
DNS-based Service Discovery methods the Pledge contacts a well
known vendor provided bootstrapping server by performing a DNS
lookup using a well known URI such as
"bootstrapks.vendor-example.com". The details of the URI are
vendor specific. Vendors that leverage this method on the Pledge
are responsible for providing the bootstrapks service.</t>
</list>DNS-based service discovery communicates the local proxy
IPv4 or IPv6 address and port to the Pledge. Once a proxy is
discovered the Pledge communicates with a Registrar through the
proxy using the bootstrapping protocol defined in <xref
target="ProtocolDetails"></xref>. The current DNS services returned
during each query is maintained until bootstrapping is completed. If
bootstrapping fails and the Pledge returns to the Discovery state it
picks up where it left off and continues attempting bootstrapping.
For example if the first Multicast DNS _bootstrapks._tcp.local
response doesn't work then the second and third responses are tried.
If these fail the Pledge moves on to normal DNS-based Service
Discovery.</t>
<t>Each discovery method attempted SHOULD exponentially back-off
attempts (to a maximum of one hour) to avoid overloading the network
infrastructure with discovery. The back-off timer for each method
MUST be independent of other methods. Methods SHOULD be run in
parallel to avoid head of queue problems. Once a connection to a
Registrar is established (e.g. establishment of a TLS session key)
there are expectations of more timely responses, see <xref
target="RequestAuditTokenFromRegistrar"></xref>.</t>
<t>Once all discovered services are attempted the device SHOULD
return to Multicast DNS. It should periodically retry the vendor
specific mechanisms. The Pledge may prioritize selection order as
appropriate for the anticipated environment.</t>
</section>
<section anchor="identity" title="Identity">
<t>The Pledge identifies itself during the communication protocol
handshake. If the client identity is rejected the Pledge repeats the
Discovery process using the next proxy or discovery method
available.</t>
<t>The bootstrapping protocol server is not initially authenticated.
Thus the connection is provisional and all data received is
untrusted until sufficiently validated even though it is over a TLS
connection. This is aligned with the existing provisional mode of
EST [RFC7030] during s4.1.1 "Bootstrap Distribution of CA
Certificates". See <xref target="AuditTokenResponse"></xref> for
more information about when the TLS connection authenticated is
completed.</t>
<t>All security associations established are between the new device
and the Bootstrapping server regardless of proxy operations.</t>
</section>
<section title="Request Join">
<t>The Pledge POSTs a request to join the domain to the
Bootstrapping server. This request contains a Pledge generated nonce
and informs the Bootstrapping server which imprint methods the
Pledge will accept.</t>
<t>As indicated in EST [RFC7030] the bootstrapping server MAY
redirect the client to an alternate server. This is most useful in
the case where the Pledge has resorted to a well known vendor URI
and is communicating with the vendor's Registrar directly. In this
case the Pledge has authenticated the Registrar using the local
Implicit Trust Anchor database and can therefore treat the redirect
URI as a trusted URI which can also be validated using the Implicit
Trust Anchor database. Since client authentication occurs during the
TLS handshake the bootstrapping server has sufficient information to
apply appropriate policy concerning which server to redirect to.</t>
<t>The nonce ensures the Pledge can verify that responses are
specific to this bootstrapping attempt. This minimizes the use of
global time and provides a substantial benefit for devices without a
valid clock.</t>
</section>
<section anchor="AcceptDomain" title="Imprint">
<t>The domain trust anchor is received by the Pledge during the
bootstrapping protocol methods in the form of a voucher. The goal of
the imprint state is to securely obtain a copy of this trust anchor
without involving human interaction.</t>
<t>The enrollment protocol EST <xref target="RFC7030"></xref>
details a set of non-autonomic bootstrapping methods such as:</t>
<t><list style="symbols">
<t>using the Implicit Trust Anchor database (not an autonomic
solution because the URL must be securely distributed),</t>
<t>engaging a human user to authorize the CA certificate using
out-of-band data (not an autonomic solution because the human
user is involved),</t>
<t>using a configured Explicit TA database (not an autonomic
solution because the distribution of an explicit TA database is
not autonomic),</t>
<t>and using a Certificate-Less TLS mutual authentication method
(not an autonomic solution because the distribution of symmetric
key material is not autonomic).</t>
</list>This document describes autonomic methods that MUST be
supported by the Pledge:</t>
<t><list style="hanging">
<t hangText="Audit Voucher">Audit Vouchers are obtained by a
Registrar from the MASA service and presented to the Pledge for
validation. These indicate to the Pledge that joining the domain
has been logged by a logging service.</t>
<t hangText="Ownership Voucher">Ownership Vouchers are obtained
by a Registrar from the MASA service and explicitly indicate the
owner of the Pledge. The Ownership Voucher is defined in <xref
target="I-D.ietf-netconf-zerotouch"></xref>.</t>
</list></t>
<t>Since client authentication occurs during the TLS handshake the
bootstrapping server has sufficient information to apply appropriate
policy concerning which method to use.</t>
<t>The Audit Voucher contains the domain's public key material as
provided to the MASA service by a Registrar. This provides
sufficient information to the client to complete automated
bootstrapping with the local key infrastructure. The Ownership
Voucher contains the Owner Certificate which the Pledge uses to
authenticate the TLS connection.</t>
<t>If the autonomic methods fail the Pledge returns to discovery
state and attempts bootstrapping with the next available discovered
Registrar.</t>
</section>
<section anchor="timeunknown" title="Lack of realtime clock">
<t>Many devices when bootstrapping do not have knowledge of the
current time. Mechanisms like Network Time Protocols can not be
secured until bootstrapping is complete. Therefore bootstrapping is
defined in a method that does not require knowledge of the current
time.</t>
<t>Unfortunately there are moments during bootstrapping when
certificates are verified, such as during the TLS handshake, where
validity periods are confirmed. This paradoxical "catch-22" is
resolved by the Pledge maintaining a concept of the current "window"
of presumed time validity that is continually refined throughout the
bootstrapping process as follows:</t>
<t><list style="symbols">
<t>Initially the Pledge does not know the current time.</t>
<t>During Pledge authentiation by the Registrar a realtime clock
can be used by the Registrar. This bullet expands on a closely
related issue regarding Pledge lifetimes. RFC5280 indicates that
long lived Pledge certifiates "SHOULD be assigned the
GeneralizedTime value of 99991231235959Z" [RFC5280] so the
Registrar MUST support such lifetimes and SHOULD support
ignoring Pledge lifetimes if they did not follow the RFC5280
recommendations.</t>
<t>Once the Audit Voucher is accepted the validity period of the
domainCAcert in the voucher (see <xref
target="AuditTokenResponse"></xref>) now describes a valid time
window. Any subsequent certificate validity periods checked
during RFC5280 path validation MUST occur within this
window.</t>
<t>When accepting an enrollment certificate the validity period
within the new certificate is assumed to be valid by the Pledge.
The Pledge is now willing to use this credential for client
authentication.</t>
</list></t>
<t>Once in this state the Pledge has a valid trust anchor with the
local domain and has a locally issued credential. These MAY be used
to secure distribution of more accurate time information although
specification of such a protocol is out-of-scope of this
document.</t>
<t>The nonce included in join attempts provides an alternate
mechanism for the Pledge to ensure Audit Voucher responses are
associated with a particular bootstrapping attempt. Nonceless Audit
Vouchers from the MASA server are always valid and thus time is not
needed.</t>
<t>Ownership Vouchers include time information and MUST be validated
using a realtime clock.</t>
</section>
<section title="Enrollment">
<t>As the final step of bootstrapping a Registrar helps to issue a
domain specific credential to the Pledge. For simplicity in this
document, a Registrar primarily facilitates issuing a credential by
acting as an RFC5280 Registration Authority for the Domain
Certification Authority.</t>
<t>Enrollment proceeds as described in <xref
target="RFC7030"></xref>. Authentication of the EST server is done
using the Voucher rather than the methods defined in EST.</t>
<t>Once the Audit or Ownership Voucher is received, as specified in
this document, the client has sufficient information to leverage the
existing communication channel with a Registrar to continue an EST
RFC7030 enrollment. Enrollment picks up at RFC7030 section 4.1.1.
bootstrapping where the Audit Voucher provides the "out-of-band" CA
certificate fingerprint (in this case the full CA certificate) such
that the client can now complete the TLS server authentication. At
this point the client continues with EST enrollment operations
including "CA Certificates Request", "CSR Attributes" and "Client
Certificate Request" or "Server-Side Key Generation".</t>
</section>
<section title="Being Managed">
<t>Functionality to provide generic "configuration" information is
supported. The parsing of this data and any subsequent use of the
data, for example communications with a Network Management System is
out of scope but is expected to occur after bootstrapping enrollment
is complete. This ensures that all communications with management
systems which can divulge local security information (e.g. network
topology or raw key material) is secured using the local credentials
issued during enrollment.</t>
<t>The Pledge uses bootstrapping to join only one domain. Management
by multiple domains is out-of-scope of bootstrapping. After the
device has successfully joined a domain and is being managed it is
plausible that the domain can insert credentials for other domains
depending on the device capabilities.</t>
<t>See <xref target="PostEnrollment"></xref>.</t>
</section>
</section>
<section anchor="proxybehaviour" title="Behavior of a Proxy">
<t>The role of the Proxy is to facilitate communications. The Proxy
forwards packets between the Pledge and a Registrar that has been
configured on the Proxy. The Proxy does not terminate the TLS
handshake. A Proxy is always assumed even if directly integrated into
a Registrar.</t>
<t>As a result of the Proxy Discovery process in section <xref
target="ProxyDiscovery"></xref>, the port number exposed by the proxy
does not need to be well known, or require an IANA allocation.</t>
<t>If the Proxy joins an Autonomic Control Plane (<xref
target="I-D.ietf-anima-autonomic-control-plane"></xref>) it
SHOULD use Autonomic Control Plane secured GRASP (<xref
target="I-D.ietf-anima-grasp"></xref>) to discovery the Registrar
address and port. For the IPIP encapsulation methods, the port
announced by the Proxy MUST be the same as on the registrar in order
for the proxy to remain stateless.</t>
<t>In order to permit the proxy functionality to be implemented on the
maximum variety of devices the chosen mechanism SHOULD use the minimum
amount of state on the proxy device. While many devices in the ANIMA
target space will be rather large routers, the proxy function is
likely to be implemented in the control plane CPU such a device, with
available capabilities for the proxy function similar to many class 2
IoT devices.</t>
<t>The document <xref
target="I-D.richardson-anima-state-for-joinrouter"></xref> provides a
more extensive analysis of the alternative proxy methods.</t>
<section anchor="coapconnection" title="CoAP connection to Registrar">
<t>The proxy MUST implement an IPIP (protocol 41) encapsulation
function for CoAP traffic to the configured UDP port on the
registrar. The proxy does not terminate the CoAP DTLS connection.
[[EDNOTE: The choice of CoAP as the mandatory to implement protocol
rather than HTTP maximizes code reuse on the smallest of devices.
Unfortunately this means this document will have to include the EST
over CoAP details as additional sections. The alternative is to make
'HTTPS proxy' method the mandatory to implement and provide a less
friendly environment for the smallest of devices. This is a decision
we'll have to see addressed by the broader team.]]</t>
<t>The IPIP encapsulation allows the proxy to forward traffic which
is otherwise not to be forwarded, as the traffic between New Node
and Proxy use IPv6 Link Local addresses.</t>
<t>If the Proxy device has more than one interface on which it
offers the proxy function, then it must select a unique (ACP) IP
address per interface in order so that the proxy can stateless
return the (link-local) reply packets to the correct link.</t>
</section>
<section title="HTTPS proxy connection to Registrar">
<t>The proxy SHOULD also provide one of: an IPIP encapsulation of
HTTP traffic on TCP port TBD to the registrar, or a TCP circuit
proxy that connects the Pledge to a Registrar.</t>
<t>When the Proxy provides a circuit proxy to a Registrar the
Registrar MUST accept HTTPS connections.</t>
<t>When the Proxy provides a stateless IPIP encapsulation to a
Registrar, then the Registrar will have to perform IPIP
decapsulation, remembering the originating outer IPIP source address
in order to qualify the inner link-local address. This is a kind of
encapsulation and processing which is similar in many ways to how
mobile IP works.</t>
<t>Being able to connect a TCP (HTTP) or UDP (CoAP) socket to a
link-local address with an encapsulated IPIP header requires API
extensions beyond <xref target="RFC3542"></xref> for UDP use, and
requires a form of connection latching (see section 4.1 of <xref
target="RFC5386"></xref> and all of <xref target="RFC5660"></xref>,
except that a simple IPIP tunnel is used rather than an IPsec
tunnel).</t>
</section>
</section>
<section title="Behavior of the Registrar">
<t>A Registrar listens for Pledges and determines if they can join the
domain. A Registrar obtains a Voucher from the MASA service and
delivers them to the Pledge as well as facilitating enrollment with
the domain PKI.</t>
<t>A Registrar is typically configured manually. If the Registrar
joins an Autonomic Control Plane (<xref
target="I-D.ietf-anima-autonomic-control-plane"></xref>) it MUST
use Autonomic Control Plane secured GRASP (<xref
target="I-D.ietf-anima-grasp"></xref>) to broadcast the Registrar's
address and port to potential Proxies.</t>
<t>Registrar behavior is as follows:</t>
<t></t>
<figure>
<artwork><![CDATA[Contacted by Pledge
+
|
+-------v----------+
| Entity | fail?
| Authentication +---------+
+-------+----------+ |
| |
+-------v----------+ |
| Entity | fail? |
| Authorization +--------->
+-------+----------+ |
| |
+-------v----------+ |
| Claiming the | fail? |
| Entity +--------->
+-------+----------+ |
| |
+-------v----------+ |
| Log Verification | fail? |
| +--------->
+-------+----------+ |
| |
+-------v----------+ +----v-------+
| Forward | | |
| Audit | | Reject |
| voucher + config | | Device |
| to the Entity | | |
+------------------+ +------------+]]></artwork>
<postamble>Figure 4</postamble>
</figure>
<section anchor="entityauthentication" title="Pledge Authentication">
<t>The applicable authentication methods detailed in EST [RFC7030]
are:</t>
<t><list style="symbols">
<t>the use of an IDevID X.509 credential during the TLS client
authentication,</t>
<t>or the use of a secret that is transmitted out of band
between the Pledge and a Registrar (this use case is not
autonomic).</t>
</list>In order to validate the IDevID X.509 credential a
Registrar maintains a database of vendor trust anchors (e.g. vendor
root certificates or keyIdentifiers for vendor root public keys).
For user interface purposes this database can be mapped to
colloquial vendor names. Registrars can be shipped with the trust
anchors of a significant number of third-party vendors within the
target market.</t>
</section>
<section anchor="AcceptingTheEntity" title="Pledge Authorization">
<t>In a fully automated network all devices must be securely
identified and authorized to join the domain.</t>
<t>A Registrar accepts or declines a request to join the domain,
based on the authenticated identity presented. Automated acceptance
criteria include:</t>
<t><list style="symbols">
<t>allow any device of a specific type (as determined by the
X.509 IDevID),</t>
<t>allow any device from a specific vendor (as determined by the
X.509 IDevID),</t>
<t>allow a specific device from a vendor (as determined by the
X.509 IDevID) against a domain white list. (The mechanism for
checking a shared white list potentiatlly used by multiple
Registrars is out of scope).</t>
</list></t>
<t>To look the Pledge up in a domain white list a consistent method
for extracting device identity from the X.509 certificate is
required. RFC6125 describes Domain-Based Application Service
identity but here we require Vendor Device-Based identity. The
subject field's DN encoding MUST include the "serialNumber"
attribute with the device's unique serial number. In the language of
RFC6125 this provides for a SERIALNUM-ID category of identifier that
can be included in a certificate and therefore that can also be used
for matching purposes. The SERIALNUM-ID whitelist is collated
according to vendor trust anchor since serial numbers are not
globally unique.</t>
<t>Since all Pledges accept Audit Vouchers a Registrar MUST use the
vendor provided MASA service to verify that the device's history log
does not include unexpected Registrars. If a device had previously
registered with another domain, a Registrar of that domain would
show in the log.</t>
<t>If a Pledge is accepted into the domain, it is expected to
request a domain certificate through a certificate enrollment
process. The result is a common trust anchor and device certificates
for all autonomic devices in a domain (these certificates can be
used for other methods, for example boundary detection,
auto-securing protocols, etc.). The authorization performed during
this phase is used for EST enrollment requests.</t>
</section>
<section title="Claiming the New Entity">
<t>Claiming an entity establishes an audit log at the MASA server
and provides a Registrar with proof, in the form of a MASA Audit
Voucher, that the log entry has been inserted. As indicated in <xref
target="AcceptDomain"></xref> a Pledge will only proceed with
bootstrapping if a validated MASA Audit Voucher has been received.
The Pledge therefore enforces that bootstrapping only occurs if the
claim has been logged. There is no requirement for the vendor to
definitively know that the device is owned by the Registrar.</t>
<t>Registrar's obtain the Vendor URI via static configuration or by
extracting it from the X.509 IDevID credential. The imprint method
supported by the Pledge is known from the X.509 IDevID credential.
[[EDNOTE: An appropriate extension for indicating the Vendor URI and
imprint method could be defined using the methods described in <xref
target="I-D.lear-mud-framework"></xref>]].</t>
<t>During initial bootstrapping the Pledge provides a nonce specific
to the particular bootstrapping attempt. The Registrar SHOULD
include this nonce when claiming the Pledge from the MASA service.
Claims from an unauthenticated Registrar are only serviced by the
MASA resource if a nonce is provided.</t>
<t>The Registrar can claim a Pledge that is not online by forming
the request using the entities unique identifier and not including a
nonce in the claim request. Audit Voucher obtained in this way do
not have a lifetime and they provide a permanent method for the
domain to claim the device. Evidence of such a claim is provided in
the audit log entries available to any future Registrar. Such claims
reduce the ability for future domains to secure bootstrapping and
therefore the Registrar MUST be authenticated by the MASA service
although no requirement is implied that the MASA associates this
authentication with ownership.</t>
<t>An Ownership Voucher requires the vendor to definitively know
that a device is owned by a specific domain. The method used to
"claim" this are out-of-scope. A MASA ignores or reports failures
when an attempt is made to claim a device that has a an Ownership
Voucher.</t>
</section>
<section title="Log Verification">
<t>A Registrar requests the log information for the Pledge from the
MASA service. The log is verified to confirm that the following is
true to the satisfaction of a Registrar's configured policy:</t>
<t><list style="symbols">
<t>Any nonceless entries in the log are associated with
domainIDs recognized by the registrar.</t>
<t>Any nonce'd entries are older than when the domain is known
to have physical possession of the Pledge or that the domainIDs
are recognized by the registrar.</t>
</list>If any of these criteria are unacceptable to a Registrar
the entity is rejected. A Registrar MAY be configured to ignore the
history of the device but it is RECOMMENDED that this only be
configured if hardware assisted NEA [RFC5209] is supported.</t>
<t>This document specifies a simple log format as provided by the
MASA service to the registar. This format could be improved by
distributed consensus technologies that integrate the Audit Voucher
with a current technologies such as block-chain or hash trees or the
like. Doing so is out of the scope of this document but are
anticipated improvements for future work.</t>
</section>
</section>
<section title="Behavior of the MASA Service">
<t>The MASA service is provided by the Factory provider on the global
Internet. The URI of this service is well known. The URI SHOULD also
be provided as an X.509 IDevID extension (a "MASA Audit Voucher
Distribution Point" extension).</t>
<t>The MASA service provides the following functionalities to
Registrars:</t>
<section title="Issue Audit Voucher and Log the event">
<t>A Registrar POSTs a claim message optionally containing the
bootstrap nonce to the MASA server.</t>
<t>If a nonce is provided the MASA service responds to all requests.
The MASA service verifies the Registrar is representative of the
domain and generates a privacy protected log entry before responding
with the Audit Voucher. For the simple log format defined in this
document using the DomainID is considered sufficient privacy. Future
work to improve the logging mechanism could include additional
privacy protections.</t>
<t>If a nonce is not provided then the MASA service MUST
authenticate the Registrar as a valid customer. This prevents denial
of service attacks.</t>
</section>
<section title="Retrieve Audit Entries from Log">
<t>When determining if a Pledge should be accepted into a domain the
Registrar retrieves a copy of the audit log from the MASA service.
This contains a list of privacy protected domain identities that
have previously claimed the device. Included in the list is an
indication of the time the entry was made and if the nonce was
included.</t>
</section>
</section>
<section anchor="PostEnrollment"
title="Leveraging the new key infrastructure / next steps">
<t>As the devices have a common trust anchor, device identity can be
securely established, making it possible to automatically deploy
services across the domain in a secure manner.</t>
<t>Examples of services:<list style="symbols">
<t>Device management.</t>
<t>Routing authentication.</t>
<t>Service discovery.</t>
</list></t>
<section anchor="boundary" title="Network boundaries">
<t>When a device has joined the domain, it can validate the domain
membership of other devices. This makes it possible to create trust
boundaries where domain members have higher level of trusted than
external devices. Using the autonomic User Interface, specific
devices can be grouped into to sub domains and specific trust levels
can be implemented between those.</t>
</section>
</section>
<section title="Interactions with Network Access Control">
<t>The assumption is that Network Access Control (NAC) completes using
the Pledge 's X.509 IDevID credentials and results in the device
having sufficient connectivity to discovery and communicate with the
proxy. Any additional connectivity or quarantine behavior by the NAC
infrastructure is out-of-scope. After the devices has completed
bootstrapping the mechanism to trigger NAC to re-authenticate the
device and provide updated network privileges is also
out-of-scope.</t>
<t>This achieves the goal of a bootstrap architecture that can
integrate with NAC but does not require NAC within the network where
it wasn't previously required. Future optimizations can be achieved by
integrating the bootstrapping protocol directly into an initial EAP
exchange.</t>
</section>
</section>
<section title="Domain Operator Activities">
<t>This section describes how an operator interacts with a domain that
supports the bootstrapping as described in this document.</t>
<section title="Instantiating the Domain Certification Authority">
<t>This is a one time step by the domain administrator. This is an
"off the shelf" CA with the exception that it is designed to work as
an integrated part of the security solution. This precludes the use of
3rd party certification authority services that do not provide support
for delegation of certificate issuance decisions to a domain managed
Registration Authority.</t>
</section>
<section title="Instantiating the Registrar">
<t>This is a one time step by the domain administrator. One or more
devices in the domain are configured take on a Registrar function.</t>
<t>A device can be configured to act as a Registrar or a device can
auto-select itself to take on this function, using a detection
mechanism to resolve potential conflicts and setup communication with
the Domain Certification Authority. Automated Registrar selection is
outside scope for this document.</t>
</section>
<section anchor="accepting" title="Accepting New Entities">
<t>For each Pledge the Registrar is informed of the unique identifier
(e.g. serial number) along with the manufacturer's identifying
information (e.g. manufacturer root certificate). This can happen in
different ways:</t>
<t><list style="numbers">
<t>Default acceptance: In the simplest case, the new device
asserts its unique identity to a Registrar. The registrar accepts
all devices without authorization checks. This mode does not
provide security against intruders and is not recommended.</t>
<t>Per device acceptance: The new device asserts its unique
identity to a Registrar. A non-technical human validates the
identity, for example by comparing the identity displayed by the
registrar (for example using a smartphone app) with the identity
shown on the packaging of the device. Acceptance may be triggered
by a click on a smartphone app "accept this device", or by other
forms of pairing. See also <xref
target="I-D.behringer-homenet-trust-bootstrap"></xref> for how the
approach could work in a homenet.</t>
<t>Whitelist acceptance: In larger networks, neither of the
previous approaches is acceptable. Default acceptance is not
secure, and a manual per device methods do not scale. Here, the
registrar is provided a priori with a list of identifiers of
devices that belong to the network. This list can be extracted
from an inventory database, or sales records. If a device is
detected that is not on the list of known devices, it can still be
manually accepted using the per device acceptance methods.</t>
<t>Automated Whitelist: an automated process that builds the
necessary whitelists and inserts them into the larger network
domain infrastructure is plausible. Once set up, no human
intervention is required in this process. Defining the exact
mechanisms for this is out of scope although the registrar
authorization checks is identified as the logical integration
point of any future work in this area.</t>
</list></t>
<t>None of these approaches require the network to have permanent
Internet connectivity. Even when the Internet based MASA service is
used, it is possible to pre-fetch the required information from the
MASA a priori, for example at time of purchase such that devices can
enroll later. This supports use cases where the domain network may be
entirely isolated during device deployment.</t>
<t>Additional policy can be stored for future authorization decisions.
For example an expected deployment time window or that a certain Proxy
must be used.</t>
</section>
<section title="Automatic Enrollment of Devices">
<t>The approach outlined in this document provides a secure zero-touch
method to enroll new devices without any pre-staged configuration. New
devices communicate with already enrolled devices of the domain, which
proxy between the new device and a Registrar. As a result of this
completely automatic operation, all devices obtain a domain based
certificate.</t>
</section>
<section title="Secure Network Operations">
<t>The certificate installed in the previous step can be used for all
subsequent operations. For example, to determine the boundaries of the
domain: If a neighbor has a certificate from the same trust anchor it
can be assumed "inside" the same organization; if not, as outside. See
also <xref target="boundary"></xref>. The certificate can also be used
to securely establish a connection between devices and central control
functions. Also autonomic transactions can use the domain certificates
to authenticate and/or encrypt direct interactions between devices.
The usage of the domain certificates is outside scope for this
document.</t>
</section>
</section>
<section anchor="ProtocolDetails" title="Protocol Details">
<t>A bootstrapping protocol could be implemented as an independent
protocol from EST, but for simplicity and to reduce the number of TLS
connections and crypto operations required on the Pledge, it is
described specifically as extensions to EST. These extensions MUST be
supported by the Registrar EST server within the same .well-known URI
tree as the existing EST URIs as described in [RFC7030] section
3.2.2.</t>
<t>The Pledge establishes a TLS connection with the Registrar through
the circuit proxy (see <xref target="proxybehaviour"></xref>) but the
TLS connection is with the Registar; so for this section the "Pledge" is
the TLS client and the "Registrar" is the TLS server.</t>
<t>Establishment of the TLS connection for bootstrapping is as specified
for EST [RFC7030]. In particular server identity and client identity are
as described in EST [RFC7030] section 3.3. In EST [RFC7030] provisional
server authentication for bootstrapping is described in section 4.1.1
wherein EST clients can "engage a human user to authorize the CA
certificate using out-of-band data such as a CA certificate" or wherein
a human user configures the URI of the EST server for Implicit TA based
authentication. As described in this document, <xref
target="CompletingAuthenticationBootstrapping"></xref>, a new method of
bootstrapping now provides a completely automating method of
bootstrapping PKI.</t>
<t>The extensions for the Pledge client are as follows:</t>
<t><list style="symbols">
<t>The Pledge provisionally accept the EST server certificate during
the TLS handshake as detailed in <xref
target="CompletingAuthenticationBootstrapping"></xref>.</t>
<t>The Pledge requests and validates the Audit Voucher as described
below. At this point the Pledge has sufficient information to
validate domain credentials.</t>
<t>The Pledge calls the EST defined /cacerts method to obtain the
current CA certificate. These are validated using the Audit
Voucher.</t>
<t>The Pledge completes bootstrapping as detailed in EST section
4.1.1.</t>
</list></t>
<t>In order to obtain a validated Audit Voucher and Audit Log a
Registrar contacts the MASA service Service using REST calls:</t>
<figure>
<artwork><![CDATA[ +-----------+ +----------+ +-----------+ +----------+
| New | | Circuit | | | | |
| Entity | | Proxy | | Registrar | | Vendor |
| | | | | | | |
++----------+ +--+-------+ +-----+-----+ +--------+-+
| | | |
| | | |
| TLS hello | TLS hello | |
Establish +---------------C---------------> |
TLS | | | |
connection | | Server Cert | |
<---------------C---------------+ |
| Client Cert | | |
+---------------C---------------> |
| | | |
HTTP REST | POST /requestvoucher | |
Data +--------------------nonce------> |
| . | /requestvoucher|
| . +---------------->
| <----------------+
| | /requestlog |
| +---------------->
| voucher <----------------+
<-------------------------------+ |
| (optional config information) | |
| . | |
| . | |]]></artwork>
<postamble>Figure 5</postamble>
</figure>
<t>In some use cases the Registrar may need to contact the Vendor in
advanced, for example when the target network is air-gapped. The
nonceless request format is provided for this and the resulting flow is
slightly different. The security differences associated with not knowing
the nonce are discussed below:</t>
<figure>
<artwork><![CDATA[ +-----------+ +----------+ +-----------+ +----------+
| New | | Circuit | | | | |
| Entity | | Proxy | | Registrar | | Vendor |
| | | | | | | |
++----------+ +--+-------+ +-----+-----+ +--------+-+
| | | |
| | | |
| | | /requestvoucher|
| | (nonce +---------------->
| | unknown) <----------------+
| | | /requestlog |
| | +---------------->
| | <----------------+
| TLS hello | TLS hello | |
Establish +---------------C---------------> |
TLS | | | |
connection | | Server Cert | |
<---------------C---------------+ |
| Client Cert | | |
| | | |
HTTP REST | POST /requestvoucher | |
Data +----------------------nonce----> (discard |
| voucher | nonce) |
<-------------------------------+ |
| (optional config information) | |
| . | |
| . | |]]></artwork>
<postamble>Figure 6</postamble>
</figure>
<t></t>
<t>The extensions for a Registrar server are as follows:</t>
<t><list style="symbols">
<t>The Registrar requests and validates the Audit Voucher from the
vendor authorized MASA service.</t>
<t>The Registrar forwards the Audit Voucher to the Pledge when
requested.</t>
<t>The Registar performs log verifications in addition to local
authorization checks before accepting the Pledge device.</t>
</list></t>
<section anchor="RequestAuditTokenFromRegistrar"
title="Request Voucher from the Registrar">
<t>When the Pledge bootstraps it makes a request for a Voucher from a
Registrar.</t>
<t>This is done with an HTTPS POST using the operation path value of
"/requestvoucher".</t>
<t>The request format is JSON object containing a 64bit nonce
generated by the client for each request. This nonce MUST be a
cryptographically strong random or pseudo-random number that can not
be easily predicted. The nonce MUST NOT be reused for multiple
attempts to join a network domain. The nonce assures the Pledge that
the Audit Voucher response is associated with this bootstrapping
attempt and is not a replay.</t>
<t>Request media type: application/auditnonce</t>
<t>Request format: a JSON file with the following:</t>
<figure>
<artwork><![CDATA[{
"version":"1",
"nonce":"<64bit nonce value>",
}]]></artwork>
</figure>
<t></t>
<t>[[EDNOTE: Even if the nonce was signed it would provide no defense
against rogue registrars; although it would assure the MASA that a
certified Pledge exists. To protect against rogue registrars a nonce
component generated by the MASA (a new round trip) would be required).
Instead this is addressed by requiring MASA & Registrar
authentications but it is worth exploring additional protections. This
to be explored more at IETF96.]]</t>
<t>The Registrar validates the client identity as described in EST
[RFC7030] section 3.3.2. The registrar performs authorization as
detailed in Section 3.3.2. If authorization is successful the
Registrar obtains an Voucher from the MASA service (see Section
5.2).</t>
<t>The received Voucher is forwarded to the Pledge.</t>
<t>As indicated in EST [RFC7030] the bootstrapping server can redirect
the client to an alternate server. If the Pledge authenticated a
Registrar using the well known URI method then the Pledge MUST follow
the redirect automatically and authenticate the new Registrar against
the redirect URI provided. If the Pledge had not yet authenticated a
Registrar because it was discovered and was not a known-to-be-valid
URI then the new Registrar must be authenticated using one of the two
autonomic methods described in this document. Similarly the Registar
MAY respond with an HTTP 202 ("the request has been accepted for
processing, but the processing has not been completed") as described
in EST [RFC7030] section 4.2.3.</t>
<t>Recall that during this communication with the Registar the TLS
authentication is only provisional. The Pledge client MUST handle all
data from the Registrar with upmost care. In particular the Pledge
MUST only allow a single redirection and MUST only support a delay of
five seconds before declaring the Registrar a failure and moving on to
the next discovered Registrar. As detailed in <xref
target="ProxyDiscovery"></xref> if no suitable Registrar is found the
Pledge restarts the state machine and tries again. So a Registrar that
is unable to complete the transaction the first time will have future
chances.</t>
</section>
<section anchor="RequestAuditToken" title="Request Voucher from MASA">
<t>A Registrar requests a Voucher from the MASA service using a REST
interface. For simplicity this is defined as an optional EST message
between a Registrar and an EST server running on the MASA service
although the Registrar is not required to make use of any other EST
functionality when communicating with the MASA service. (The MASA
service MUST properly reject any EST functionality requests it does
not wish to service; a requirement that holds for any REST
interface).</t>
<t>This is done with an HTTP POST using the operation path value of
"/requestvoucher".</t>
<t>The request format is a JSON object optionally containing the nonce
value (as obtained from the bootstrap request) and the X.509 IDevID
extracted serial number (the full certificate is not needed and no
proof-of-possession information for the device identity is included).
The AuthorityKeyIdentifier value from the certificate is included to
ensure a statistically unique identity. The Pledge's serial number is
extracted from the X.509 IDevID subject name id-at-serialNumber or it
is the base64 encoded RFC4108 hardwareModuleName hwSerialNum:</t>
<figure>
<artwork><![CDATA[{
"version":"1",
"nonce":"<64bit nonce value>",
"IDevIDAuthorityKeyIdentifier":"<base64 encoded keyIdentifier">,
"DevIDSerialNumber":"<id-at-serialNumber or base64 encoded
hardwareModuleName hwSerialNum>",
}]]></artwork>
</figure>
<t></t>
<t>A Registrar MAY exclude the nonce from the request. Doing so allows
the Registrar to request a Voucher when the Pledge is not online, or
when the target bootstrapping environment is not on the same network
as the MASA server (this requires the Registrar to learn the
appropriate DevIDSerialNumber field from the physical device labeling
or from the sales channel -- how this occurs is out-of-scope of this
document). If a nonce is not provided the MASA server MUST
authenticate the client as described in EST [RFC7030] section 3.3.2 to
reduce the risk of DDoS attacks. A Registrar performs authorization as
detailed in <xref target="AcceptingTheEntity"></xref>. If
authorization is successful the Registrar obtains an Voucher from the
MASA service (see <xref target="RequestAuditToken"></xref>).</t>
<t>The JSON message information is encapsulated in a <xref
target="RFC5652"></xref> Signed-data that is signed by the Registrar.
The entire certificate chain, up to and including the Domain CA, MUST
be included in the CertificateSet structure. The MASA service checks
the internal consistency of the CMS but does not authenticate the
domain identity information. The domain is not know to the MASA server
in advance and a shared trust anchor is not implied. The MASA server
MUST verify that the CMS is signed by a Registrar certificate (by
checking for the cmc-idRA field) that was issued by a the root
certificate included in the CMS. This ensures that the Registrar
making the claim is an authorized Registrar of the unauthenticated
domain. The EST style client authentication (TLS and HTTP) is used to
provide a DDoS prevention strategy.</t>
<t>The root certificate is extracted and used to populate the Audit
Voucher. The domain ID (e.g. hash of the public key of the domain) is
extracted from the root certificate and is used to update the audit
log.</t>
</section>
<section anchor="AuditTokenResponse" title="Audit Voucher Response">
<t>The voucher response to requests from the device and requests from
a Registrar are in the same format. A Registrar either caches prior
MASA responses or dynamically requests a new Voucher based on local
policy.</t>
<t>If the the join operation is successful, the server response MUST
contain an HTTP 200 response code with a content-type of
"application/authorizationvoucher". The server MUST answer with a
suitable 4xx or 5xx HTTP [RFC2616] error code when a problem occurs.
The response data from the MASA server MUST be a plaintext
human-readable error message containing explanatory information
describing why the request was rejected.</t>
<t>The Audit Voucher consists of the nonce, if supplied, the serial
number information identifying the device and the domain CA
certificate extracted from the request:</t>
<figure>
<artwork><![CDATA[{
"version":"1",
"nonce":"<64bit nonce value>",
"IDevIDAuthorityKeyIdentifier":"<base64 encoded keyIdentifier>",
"DevIDSerialNumber":"<id-at-serialNumber>",
"domainCAcert":"<the base64 encoded domain CA's certificate>"
}]]></artwork>
</figure>
<t>The Audit Voucher response is encapsulated in a <xref
target="RFC5652"></xref> Signed-data that is signed by the MASA
server. The Pledge verifies this signed message using the manufacturer
installed trust anchor assocaited with the X.509 IDevID. [[EDNOTE: As
detailed in netconf-zerotouch this might be a distinct trust anchor
rather than re-using the trust anchor for the IDevID. This concept
will need to be detailed in this document as well.]]</t>
<t>[[EDNOTE: Using CMS is consistent with the alignment of this
bootstrapping document with EST, a PKIX enrollment protocol that
includes Certificate Management over CMS. An alternative format would
be the RFC7515 JSON Web Signature (JWS), which would allow clients
that do not use fullCMC messages to avoid CMS entirely. Use of JWS
would likely include a discussion of CBOR in order ensure the base64
expansions of the certs and signatures within the JWS message are of
minimal size -- it is not yet clear to this author how that would work
out]]</t>
<t>The 'domainCAcert' element of this message contains the domain CA's
public key. This is specific to bootstrapping a public key
infrastructure. To support bootstrapping other key infrastructures
additional domain identity types might be defined in the future.
Clients MUST be prepared to ignore additional fields they do not
recognize. Clients MUST be prepared to parse and fail gracefully from
an Audit Voucher response that does not contain a 'domainCAcert' field
at all.</t>
<t>To minimize the size of the Audit Voucher response message the
domainCAcert is not a complete distribution of the EST section 4.1.3
CA Certificate Response.</t>
<t>The Pledge installs the domainCAcert trust anchor. As indicated in
<xref target="identity"></xref> the newly installed trust anchor is
used as an EST RFC7030 Explicit Trust Anchor. The Pledge MUST use the
domainCAcert trust anchor to immediately validate the currently
provisional TLS connection to a Registrar.</t>
<section anchor="CompletingAuthenticationBootstrapping"
title="Completing authentication of Provisional TLS connection">
<t>If a Registrar's credential can not be verified using the
domainCAcert trust anchor the TLS connection is immediately
discarded and the Pledge abandons attempts to bootstrap with this
discovered registrar.</t>
<t>The following behaviors on a Registrar and Pledge are in addition
to normal PKIX operations:</t>
<t><list style="symbols">
<t>The EST server MUST use a certificate that chains to the
domainCAcert. This means that when the EST server obtains
renewed credentials the credentials included in the <xref
target="RequestAuditToken"></xref> request match the chain used
in the current provisional TLS connection.</t>
<t>The Pledge PKIX path validation of a Registrar validity
period information is as described in <xref
target="timeunknown"></xref>.</t>
</list>Because the domainCAcert trust anchor is installed as an
Explicit Trust Anchor it can be used to authenticate any dynamically
discovered EST server that contain the id-kp-cmcRA extended key
usage extension as detailed in EST RFC7030 section 3.6.1; but to
reduce system complexity the Pledge SHOULD avoid additional
discovery operations. Instead the Pledge SHOULD communicate directly
with the Registrar as the EST server to complete PKI local
certificate enrollment. Additionally the Pledge SHOULD use the
existing TLS connection to proceed with EST enrollment, thus
reducing the total amount of cryptographic and round trip operations
required during bootstrapping. [[EDNOTE: It is reasonable to mandate
that the existing TLS connection be re-used? e.g. MUST >>
SHOULD?]]</t>
</section>
</section>
<section title="Voucher Status Telemetry ">
<t>For automated bootstrapping of devices the adminstrative elements
providing bootstrapping also provide indications to the system
administrators concerning device lifecycle status. To facilitate this
those elements need telemetry information concerning the device's
status.</t>
<t>To indicate Pledge status regarding the Audit Voucher the client
SHOULD post a status message.</t>
<t>The client HTTP POSTs the following to the server at the EST well
known URI /voucher_status. The Status field indicates if the Voucher
was acceptable. If it was not acceptable the Reason string indicates
why. In the failure case this message is being sent to an
unauthenticated, potentially malicious Registrar and therefore the
Reason string SHOULD NOT provide information beneficial to an
attacker. The operational benefit of this telemetry information is
balanced against the operational costs of not recording that an
Voucher was ignored by a client the registar expected to continue
joining the domain.</t>
<t><figure>
<artwork><![CDATA[{
"version":"1",
"Status":FALSE /* TRUE=Success, FALSE=Fail"
"Reason":"Informative human readable message"
}]]></artwork>
</figure>The server SHOULD respond with an HTTP 200 but MAY simply
fail with an HTTP 404 error. The client ignores any response. Within
the server logs the server SHOULD capture this telemetry
information.</t>
</section>
<section title="MASA authorization log Request ">
<t>A registrar requests the MASA authorization log from the MASA
service using this EST extension.</t>
<t>This is done with an HTTP GET using the operation path value of
"/requestauditlog".</t>
<t>The client HTTP POSTs the same Voucher Request as for requesting an
audit token but now posts it to the /requestauditlog URI instead. The
IDevIDAuthorityKeyIdentifier and DevIDSerialNumber informs the MASA
server which log is requested so the appropriate log can be prepared
for the response.</t>
</section>
<section title="MASA authorization log Response">
<t>A log data file is returned consisting of all log entries. For
example:</t>
<t><figure>
<artwork><![CDATA[{
"version":"1",
"events":[
{
"date":"<date/time of the entry>",
"domainID":"<domainID as extracted from the domain CA certificate
within the CMS of the audit voucher request>",
"nonce":"<any nonce if supplied (or the exact string 'NULL')>"
},
{
"date":"<date/time of the entry>",
"domainID":"<domainID as extracted from the domain CA certificate
within the CMS of the audit voucher request>",
"nonce":"<any nonce if supplied (or the exact string 'NULL')>"
}
]
}]]></artwork>
</figure></t>
<t>Distribution of a large log is less than ideal. This structure can
be optimized as follows: All nonce-less entries for the same domainID
MAY be condensed into the single most recent nonceless entry.</t>
<t>A Registrar uses this log information to make an informed decision
regarding the continued bootstrapping of the Pledge. For example if
the log includes unexpected domainIDs this is indicative of
problematic imprints by the Pledge. If the log includes nonce-less
entries this is indicative of the permanent ability for the indicated
domain to trigger a reset of the device and take over management of
it. Equipment that is purchased pre-owned can be expected to have an
extensive history.</t>
<t>Log entries containing the Domain's ID can be compared against
local history logs in search of discrepancies.</t>
</section>
<section title="EST Integration for PKI bootstrapping">
<t>The prior sections describe EST extensions necessary to enable
fully automated bootstrapping. Although the Audit Voucher
request/response structure members IDevIDAuthorityKeyIdentifier and
DevIDSerialNumber are specific to PKI bootstrapping these are the only
PKI specific aspects of the extensions and future work might replace
them with non-PKI structures.</t>
<t>The prior sections provide functionality for the Pledge to obtain a
trust anchor representative of the Domain. The following section
describe using EST to obtain a locally issued PKI certificate. The
Pledge SHOULD leverage the discovered Registrar to proceed with
certificate enrollment and, if they do, MUST implement the EST options
described in this section. The Pledge MAY perform alternative
enrollment methods including discovering an alternate EST server, or
proceed to use its IDevID credential indefinitely.</t>
<section title="EST Distribution of CA Certificates">
<t>The Pledge MUST request the full EST Distribution of CA
Certificates message. See RFC7030, section 4.1.</t>
<t>This ensures that the Pledge has the complete set of current CA
certificates beyond the domainCAcert (see <xref
target="AuditTokenResponse"></xref> for a discussion of the
limitations). Although these restrictions are acceptable for a
Registrar integrated with initial bootstrapping they are not
appropriate for ongoing PKIX end entity certificate validation.</t>
</section>
<section title="EST CSR Attributes">
<t>Automated bootstrapping occurs without local administrative
configuration of the Pledge. In some deployments its plausible that
the Pledge generates a certificate request containing only identity
information known to the Pledge (essentially the IDevID information)
and ultimately receives a certificate containing domain specific
identity information. Conceptually the CA has complete control over
all fields issued in the end entity certificate. Realistically this
is operationally difficult with the current status of PKI
certificate authority deployments where the CSR is submitted to the
CA via a number of non-standard protocols.</t>
<t>To alleviate operational difficulty the Pledge MUST request the
EST "CSR Attributes" from the EST server. This allows the local
infrastructure to inform the Pledge of the proper fields to include
in the generated CSR.</t>
<t>[[EDNOTE: The following is specific to anima purposes and should
be moved to an appropriate anima document so as to keep
bootstrapping as generic as possible: What we want are a 'domain
name' stored in [TBD] and an 'ACP IPv6 address' stored in the
iPAddress field as specified in RFC5208 s4.2.1.6. ref ACP draft
where certificate verification [TBD]. These should go into the
subjectaltname in the [TBD] fields.]]. If the hardwareModuleName in
the IDevID is populated then it SHOULD by default be propagated to
the LDevID along with the hwSerialNum. The registar SHOULD support
local policy concerning this functionality. [[EDNOTE: extensive use
of EST CSR Attributes might need an new OID definition]].]]</t>
<t>The Registar MUST also confirm the resulting CSR is formatted as
indicated before forwarding the request to a CA. If the Registar is
communicating with the CA using a protocol like full CMC which
provides mechanisms to override the CSR attributes, then these
mechanisms MAY be used even if the client ignores CSR Attribute
guidance.</t>
</section>
<section title="EST Client Certificate Request">
<t>The Pledge MUST request a new client certificate. See RFC7030,
section 4.2.</t>
</section>
<section title="Enrollment Status Telemetry">
<t>For automated bootstrapping of devices the adminstrative elements
providing bootstrapping also provide indications to the system
administrators concerning device lifecycle status. This might
include information concerning attempted bootstrapping messages seen
by the client, MASA provides logs and status of credential
enrollment. The EST protocol assumes an end user and therefore does
not include a final success indication back to the server. This is
insufficient for automated use cases.</t>
<t>To indicate successful enrollment the client SHOULD re-negotiate
the EST TLS session using the newly obtained credentials. This
occurs by the client initiating a new TLS ClientHello message on the
existing TLS connection. The client MAY simply close the old TLS
session and start a new one. The server MUST support either
model.</t>
<t>In the case of a failure the Reason string indicates why the most
recent enrollment failed. The SubjectKeyIdentifier field MUST be
included if the enrollment attempt was for a keypair that is locally
known to the client. If EST /serverkeygen was used and failed then
the this field is ommited from the status telemetry.</t>
<t>The client HTTP POSTs the following to the server at the new EST
well known URI /enrollstatus.</t>
<t><figure>
<artwork><![CDATA[{
"version":"1",
"Status":TRUE /* TRUE=Success, FALSE=Fail"
"Reason":"Informative human readable message"
"SubjectKeyIdentifier":"<base64 encoded subjectkeyidentifier for the
enrollment that failed>"
}]]></artwork>
</figure></t>
<t>The server SHOULD respond with an HTTP 200 but MAY simply fail
with an HTTP 404 error.</t>
<t>Within the server logs the server MUST capture if this message
was recieved over an TLS session with a matching client certificate.
This allows for clients that wish to minimize their crypto
operations to simpy POST this response without renegotiating the TLS
session - at the cost of the server not being able to accurately
verify that enrollment was truly successful.</t>
</section>
<section title="EST over CoAP">
<t>[[EDNOTE: In order to support smaller devices the above section
on Proxy behavior introduces mandatory to implement support for CoAP
support by the Proxy. This implies similar support by the Pledge and
Registrar and means that the EST protocol operation encapsulation
into CoAP needs to be described. EST is HTTP based and "CoaP is
designed to easily interface with HTTP for integration" [RFC7252].
Use of CoAP implies Datagram TLS (DTLS) wherever this document
describes TLS handshake specifics. A complexity is that the large
message sizes necessary for bootstrapping will require support for
[draft-ietf-core-block].]]</t>
</section>
</section>
</section>
<section anchor="reducedsecuritymodes"
title="Reduced security operational modes">
<t>A common requirement of bootstrapping is to support less secure
operational modes for support specific use cases. The following sections
detail specific ways that the Pledge, Registrar and MASA can be
configured to run in a less secure mode for the indicated reasons.</t>
<section title="Trust Model">
<figure>
<artwork><![CDATA[
+--------+ +---------+ +------------+ +------------+
| New | | Circuit | | Domain | | Vendor |
| Entity | | Proxy | | Registrar | | Service |
| | | | | | | (Internet |
+--------+ +---------+ +------------+ +------------+
]]></artwork>
<postamble>Figure 7</postamble>
</figure>
<t><list style="hanging">
<t hangText="Pledge:">The Pledge could be compromised and
providing an attack vector for malware. The entity is trusted to
only imprint using secure methods described in this document.
Additional endpoint assessment techniques are RECOMMENDED but are
out-of-scope of this document.</t>
<t hangText="Proxy:">Provides proxy functionalities but is not
involved in security considerations.</t>
<t hangText="Registrar:">When interacting with a MASA server a
Registrar makes all decisions. When Ownership Vouchers are
involved a Registrar is only a conduit and all security decisions
are made on the vendor service.</t>
<t hangText="Vendor Service, MASA:">This form of vendor service is
trusted to accurately log all claim attempts and to provide
authoritative log information to Registrars. The MASA does not
know which devices are associated with which domains. These claims
could be strengthened by using cryptographic log techniques to
provide append only, cryptographic assured, publicly auditable
logs. Current text provides only for a trusted vendor.</t>
<t hangText="Vendor Service, Ownership Validation:">This form of
vendor service is trusted to accurately know which device is owned
by which domain.</t>
</list></t>
</section>
<section title="New Entity security reductions">
<t>The Pledge MAY support "trust on first use" on physical interfaces
but MUST NOT support "trust on first use" on network interfaces. This
is because "trust on first use" permanently degrades the security for
all other use cases.</t>
<t>The Pledge MAY have an operational mode where it skips Voucher
validation one time. For example if a physical button is depressed
during the bootstrapping operation. This can be useful if the vendor
service is unavailable. This behavior SHOULD be available via local
configuration or physical presence methods to ensure new entities can
always be deployed even when autonomic methods fail. This allows for
unsecured imprint.</t>
<t>It is RECOMMENDED that this only be available if hardware assisted
NEA [RFC5209] is supported.</t>
</section>
<section title="Registrar security reductions">
<t>A Registrar can choose to accept devices using less secure methods.
These methods are acceptable when low security models are needed, as
the security decisions are being made by the local administrator, but
they MUST NOT be the default behavior:<list style="numbers">
<t>A registrar MAY choose to accept all devices, or all devices of
a particular type, at the administrator's discretion. This could
occur when informing all Registrars of unique identifiers of new
entities might be operationally difficult.</t>
<t>A registrar MAY choose to accept devices that claim a unique
identity without the benefit of authenticating that claimed
identity. This could occur when the Pledge does not include an
X.509 IDevID factory installed credential. New Entities without an
IDevID credential MAY form the <xref
target="RequestAuditTokenFromRegistrar"></xref> request using the
<xref target="RequestAuditToken"></xref> format to ensure the
Pledge's serial number information is provided to the Registar
(this includes the IDevIDAuthorityKeyIdentifier value which would
be statically configured on the Pledge). The Pledge MAY refused to
provide a TLS client certificate (as one is not available). The
Pledge SHOULD support HTTP-based or certificate-less TLS
authentication as described in EST RFC7030 section 3.3.2. A
Registrar MUST NOT accept unauthenticated New Entities unless it
has been configured to do so by an administrator that has verified
that only expected new entities can communicate with a Registrar
(presumably via a physically secured perimeter).</t>
<t>A Registrar MAY request nonce-less Audit Vouchers from the MASA
service (by not including a nonce in the request). These Audit
Vouchers can then be transmitted to the Registrar and stored until
they are needed during bootstrapping operations. This is for use
cases where target network is protected by an air gap and
therefore can not contact the MASA service during Pledge
deployment.</t>
<t>A registrar MAY ignore unrecognized nonce-less Audit Log
entries. This could occur when used equipment is purchased with a
valid history being deployed in air gap networks that required
permanent Audit Vouchers.</t>
</list></t>
<t>These modes are not available for devices that require a vendor
Ownership Voucher. The methods vendors use to determine which devices
are owned by which domains is out-of-scope.</t>
</section>
<section title="MASA security reductions">
<t>Lower security modes chosen by the MASA service effect all device
deployments unless bound to the specific device identities. In which
case these modes can be provided as additional features for specific
customers. The MASA service can choose to run in less secure modes
by:</t>
<t><list style="numbers">
<t>Not enforcing that a Nonce is in the Audit Voucher. This
results in distribution of Audit Voucher that never expire and in
effect makes the Domain an always trusted entity to the Pledge
during any subsequent bootstrapping attempts. That this occurred
is captured in the log information so that the Domain registrar
can make appropriate security decisions when a Pledge joins the
Domain. This is useful to support use cases where Registrars might
not be online during actual device deployment. Because this
results in long lived Audit Voucher and do not require the proof
that the device is online this is only accepted when the Registrar
is authenticated by the MASA server and authorized to provide this
functionality. The MASA server is RECOMMENDED to use this
functionality only in concert with Ownership Validation
tracking.</t>
<t>Not verifying ownership before responding with an Audit
Voucher. This is expected to be a common operational model because
doing so relieves the vendor providing MASA services from having
to tracking ownership during shipping and supply chain and allows
for a very low overhead MASA service. A Registrar uses the audit
log information as a defense in depth strategy to ensure that this
does not occur unexpectedly (for example when purchasing new
equipment the Registrar would throw an error if any audit log
information is reported).</t>
</list></t>
</section>
</section>
<section title="Security Considerations">
<t>In order to support a wide variety of use cases, devices can be
claimed by a registrar without proving possession of the device in
question. This would result in a nonceless, and thus always valid,
claim. Or would result in an invalid nonce being associated with a
claim. The MASA service is required to authenticate such Registrars but
no programmatic method is provided to ensure good behavior by the MASA
service. Nonceless entries into the audit log therefore permanently
reduce the value of a device because future Registrars, during future
bootstrap attempts, would now have to be configured with policy to
ignore previously (and potentially unknown) domains.</t>
<t>Future registrars are recommended to take the audit history of a
device into account when deciding to join such devices into their
network. If the MASA server were to have allowed a significantly large
number of claims this might become onerous to the MASA server which must
maintain all the extra log entries. Ensuring a Registrar is
representative of a valid customer domain even without validating
ownership helps to mitigate this.</t>
<t>It is possible for an attacker to send an authorization request to
the MASA service directly after the real Registrar obtains an
authorization log. If the attacker could also force the bootstrapping
protocol to reset there is a theoretical opportunity for the attacker to
use the Audit Voucher to take control of the Pledge but then proceed to
enroll with the target domain. Possible prevention mechanisms
include:</t>
<t><list style="symbols">
<t>Per device rate limits on the MASA service ensure such timing
attacks are difficult.</t>
<t>In the advent of an unexpectedly lost bootstrapping connection
the Registrar repeats the request for audit log information.</t>
</list></t>
<t>To facilitate logging and administrative oversight the Pledge reports
on Audit Voucher parsing status to the Registrar. In the case of a
failure this information is informative to a potentially malicious
Registar but this is RECOMMENDED anyway because of the operational
benefits of an informed administrator in cases where the failure is
indicative of a problem.</t>
<t>As indicated in EST [RFC7030] the connection is provisional and
untrusted until the server is successfully authorized. If the server
provides a redirect response the client MUST follow the redirect but the
connection remains provisional. If the client uses a well known URI for
contacting a well known Registrar the EST Implicit Trust Anchor database
is used as is described in RFC6125 to authenticate the well known URI.
In this case the connection is not provisional and RFC6125 methods can
be used for each subsequent redirection.</t>
<t>To facilitate truely limited clients EST RFC7030 section 3.3.2
requirements that the client MUST support a client authentication model
have been reduced in <xref target="reducedsecuritymodes"></xref> to a
statement that clients only "SHOULD" support such a model. This reflects
current (not great) practices but is NOT RECOMMENDED.</t>
<t>The MASA service could lock a claim and refuse to issue a new voucher
or the MASA service could go offline (for example if a vendor went out
of business). This functionality provides benefits such as theft
resistance, but it also implies an operational risk to the Domain that
Vendor behavior could limit future bootstrapping of the device by the
Domain. This can be mitigated by Registrars that request nonce-less
Audit Vouchers.</t>
<section title="Security concerns with discovery process">
<section title="Discovery of Registrar by Proxy">
<t>
As described in section <xref target="proxybehaviour" />, the
RECOMMENDED mechanism is for the proxy to discover the address of
the registrar via GRASP <xref target="I-D.ietf-anima-grasp" />
</t>
<t>
GRASP is intended to run over a secured, and private Autonomic
Control Plan <xref
target="I-D.ietf-anima-autonomic-control-plane" />.
This discovery is between the already registered Registrar,
and the already registered Proxy. There are no GRASP security
issues with this part, as both entities will have already joined
the secured ACP.
</t>
</section>
<section title="Discovery of Proxy by New Entity">
<t>
[[EDNOTE: To be discussed]]
</t>
</section>
</section>
</section>
<section title="Acknowledgements">
<t>We would like to thank the various reviewers for their input, in
particular Markus Stenberg, Brian Carpenter, Fuyu Eleven, Toerless
Eckert, Eliot Lear and Sergey Kasatkin.</t>
</section>
</middle>
<back>
<references title="Normative References">
&RFC2119;
&RFC7030;
&RFC5652;
&RFC5280;
<?rfc include="reference.RFC.3542" ?>
<?rfc include="reference.RFC.5386" ?>
<?rfc include="reference.RFC.5660" ?>
<?rfc include="reference.RFC.7228" ?>
<?rfc include="reference.RFC.6762" ?>
<?rfc include="reference.RFC.6763" ?>
<?rfc include="reference.RFC.3927" ?>
<?rfc include="reference.RFC.4862" ?>
<reference anchor="IDevID"
target="http://standards.ieee.org/findstds/standard/802.1AR-2009.html">
<front>
<title>IEEE 802.1AR Secure Device Identifier</title>
<author surname="IEEE Standard"></author>
<date month="December" year="2009" />
</front>
</reference>
</references>
<references title="Informative References">
&I-D.behringer-autonomic-network-framework;
&RFC7575;
<?rfc include="reference.I-D.ietf-anima-autonomic-control-plane" ?>
<?rfc include="reference.I-D.ietf-ace-actors" ?>
<?rfc include="reference.I-D.ietf-anima-grasp" ?>
<?rfc include="reference.I-D.richardson-anima-state-for-joinrouter" ?>
<?rfc include="reference.I-D.lear-mud-framework" ?>
<?rfc include="reference.I-D.ietf-netconf-zerotouch" ?>
<reference anchor="imprinting"
target="https://en.wikipedia.org/wiki/Imprinting_(psychology)">
<front>
<title>Wikipedia article: Imprinting</title>
<author surname="Wikipedia"></author>
<date month="July" year="2015" />
</front>
</reference>
<reference anchor="Stajano99theresurrecting"
target="https://www.cl.cam.ac.uk/~fms27/papers/1999-StajanoAnd-duckling.pdf">
<front>
<title>The resurrecting duckling: security issues for ad-hoc
wireless networks</title>
<author fullname="Frank Stajano" initials="F." surname="Stajano"></author>
<author fullname="Ross Anderson" initials="R." surname="Anderson"></author>
<date year="1999" />
</front>
</reference>
<!-- http://dictionary.reference.com/cite.html?qh=pledge&ia=luna
@article {Dictionary.com2015,
title = {Dictionary.com Unabridged},
month = {Jul},
day = {03},
year = {2015},
url = {http://dictionary.reference.com/browse/pledge},
} -->
<reference anchor="pledge"
target="http://dictionary.reference.com/browse/pledge">
<front>
<title>Dictionary.com Unabridged</title>
<author surname="Dictionary.com"></author>
<date month="July" year="2015" />
</front>
</reference>
</references>
</back>
</rfc>
<!--
Local Variables:
mode: xml
End:
-->
| PAFTECH AB 2003-2026 | 2026-04-21 09:02:52 |