One document matched: draft-hansbury-sacm-oval-info-model-mapping-01.xml
<?xml version="1.0" encoding="US-ASCII"?>
<!-- This template is for creating an Internet Draft using xml2rfc,
which is available here: http://xml.resource.org. -->
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
<!-- One method to get references from the online citation libraries.
There has to be one entity for each item to be referenced.
An alternate method (rfc include) is described in the references. -->
<!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml">
<!ENTITY RFC2865 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2865.xml">
<!ENTITY RFC3444 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3444.xml">
]>
<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>
<!-- used by XSLT processors -->
<!-- For a complete list and description of processing instructions (PIs),
please see http://xml.resource.org/authoring/README.html. -->
<!-- Below are generally applicable Processing Instructions (PIs) that most I-Ds might want to use.
(Here they are set differently than their defaults in xml2rfc v1.32) -->
<?rfc strict="yes" ?>
<!-- give errors regarding ID-nits and DTD validation -->
<!-- control the table of contents (ToC) -->
<?rfc toc="yes"?>
<!-- generate a ToC -->
<?rfc tocdepth="4"?>
<!-- the number of levels of subsections in ToC. default: 3 -->
<!-- control references -->
<?rfc symrefs="yes"?>
<!-- use symbolic references tags, i.e, [RFC2119] instead of [1] -->
<?rfc sortrefs="yes" ?>
<!-- sort the reference entries alphabetically -->
<!-- control vertical white space
(using these PIs as follows is recommended by the RFC Editor) -->
<?rfc compact="yes" ?>
<!-- do not start each main section on a new page -->
<?rfc subcompact="no" ?>
<!-- keep one blank line between list items -->
<!-- end of list of popular I-D processing instructions -->
<rfc category="info"
docName="draft-hansbury-sacm-oval-info-model-mapping-01"
ipr="trust200902">
<!-- category values: std, bcp, info, exp, and historic
ipr values: full3667, noModification3667, noDerivatives3667
you can add the attributes updates="NNNN" and obsoletes="NNNN"
they will automatically be output with "(if approved)" -->
<!-- ***** FRONT MATTER ***** -->
<front>
<!-- The abbreviated title is used in the page header - it is only necessary if the
full title is longer than 39 characters -->
<title abbrev="OVAL and the SACM Information Model">OVAL and the SACM
Information Model</title>
<!-- add 'role="editor"' below for the editors if appropriate -->
<!-- Another author who claims to be an editor -->
<author fullname="Matthew Hansbury" initials="M.H." surname="Hansbury">
<organization>The MITRE Corporation</organization>
<address>
<postal>
<street>202 Burlington Road</street>
<!-- Reorder these if your country does things differently -->
<city>Bedford</city>
<region>MA</region>
<code>01730</code>
<country>USA</country>
</postal>
<phone/>
<email>mhansbury@mitre.org</email>
<!-- uri and facsimile elements may also be added -->
</address>
</author>
<author fullname="Daniel Haynes" initials="D.H." surname="Haynes">
<organization>The MITRE Corporation</organization>
<address>
<postal>
<street>202 Burlington Road</street>
<!-- Reorder these if your country does things differently -->
<city>Bedford</city>
<region>MA</region>
<code>01730</code>
<country>USA</country>
</postal>
<phone/>
<email>dhaynes@mitre.org</email>
<!-- uri and facsimile elements may also be added -->
</address>
</author>
<author fullname="Juan Gonzalez" initials="J.G." surname="Gonzalez">
<organization>Department of Homeland Security</organization>
<address>
<postal>
<street>245 Murray Lane</street>
<!-- Reorder these if your country does things differently -->
<city>Washington</city>
<region>DC</region>
<code>20548</code>
<country>USA</country>
</postal>
<phone/>
<email>juan.gonzalez@dhs.gov</email>
<!-- uri and facsimile elements may also be added -->
</address>
</author>
<date month="November" year="2015"/>
<!-- If the month and year are both specified and are the current ones, xml2rfc will fill
in the current day for you. If only the current year is specified, xml2rfc will fill
in the current day and month for you. If the year is not the current one, it is
necessary to specify at least a month (xml2rfc assumes day="1" if not specified for the
purpose of calculating the expiry date). With drafts it is normally sufficient to
specify just the year. -->
<!-- Meta-data Declarations -->
<area>Security</area>
<workgroup>Security Automation and Continuous Monitoring</workgroup>
<!-- WG name at the upperleft corner of the doc,
IETF is fine for individual submissions.
If this element is not present, the default is "Network Working Group",
which is used by the RFC Editor as a nod to the history of the IETF. -->
<keyword>security automation</keyword>
<keyword>continuous monitoring</keyword>
<keyword>endpoint</keyword>
<keyword>posture assessment</keyword>
<keyword>oval</keyword>
<keyword>lessons learned</keyword>
<keyword>gaps</keyword>
<keyword>configuration management</keyword>
<keyword>vulnerability management</keyword>
<keyword>information model</keyword>
<!-- Keywords will be incorporated into HTML output
files in a meta tag but they have no effect on text or nroff
output. If you submit your draft to the RFC Editor, the
keywords will be used for the search engine. -->
<abstract>
<t>The OVAL community has spent more than ten years developing and
employing the OVAL Language. During this time, the community has
made a number of design decisions and learned a number of
lessons that should be leveraged as next generation endpoint
posture assessment is formulated. There are a number of places
throughout the SACM Information Model document that could be
fulfilled by portions of the OVAL Language, either in its
current state or, in some cases, with modifications. Another
output of the work executed under the OVAL project is a number
of lessons that are applicable to the SACM work. These lessons
include a clear separation of data collection and evaluation; a
call to focus on ensuring both primary source vendors and third
party security experts feel invited to the discussion and are
empowered to leverage their unique domain knowledge; and to
strive for simplicity and flexibility, where possible. Finally,
the OVAL community has a set of clear recommendations with
respect to which parts of OVAL should be used by SACM as a means
to make best use of the efforts of those that have worked on and
supported OVAL over the past ten years. Those recommendations
are: <list style="symbols">
<t>Use the OVAL System Characteristics Model as a base data
model for at least one way to provide data
collection.</t>
<t>Use the OVAL Definitions Model in parts as a base data
model for both evaluation and collection guidance.</t>
<t>Do not use the OVAL Results Model for a data model to
encode evaluation results.</t>
</list>
</t>
</abstract>
</front>
<middle>
<section title="Introduction">
<t>The Security Automation and Continuous Monitoring (SACM) IETF
Working Group <xref target="SACM"/> has been chartered with standardizing the
mechanisms by which endpoint security assessment is performed.
This includes software inventory, compliance and vulnerability
management, and other related activities. The Working Group has
created a series of artifacts <xref target="SACM-DOCUMENTS"/> to capture the
important concepts required to accomplish this goal. In addition
to Use Cases, Requirements, and Architecture documents, the
Working Group has created an initial draft of an Information
Model that describes the high-level components and concepts that
fulfill the already defined requirements.</t>
<t>This white paper discusses how the Open Vulnerability and
Assessment Language (OVAL) <xref target="OVAL-WEBSITE"/> can be leveraged in
order to implement the Information Model defined by the SACM
group. This paper is not meant to suggest that the entire OVAL
Data Model could-or even should-be supported by SACM; rather, it
breaks apart the various components of the OVAL Language and
discusses how each could be used to satisfy parts of the
Information Model.</t>
<t>This document assumes that the reader is already familiar with
OVAL and its structures. For those readers that require more
in-depth information about OVAL, please review the OVAL Tutorial
documentation <xref target="OVAL-DEFINITION-TUTORIAL"/> and other related
documentation on the OVAL website. This document describes how
these structures can be thought of as data models whose scopes
and activities overlap with the SACM Information Model.</t>
<t>Additionally, in later sections, the paper presents lessons
learned from the ten plus years of OVAL development and
curation, as well as related gaps.</t>
<section title="Requirements Language">
<t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as
described in <xref target="RFC2119">RFC 2119</xref>.</t>
</section>
</section>
<!-- This PI places the pagebreak correctly (before the section title) in the text output. -->
<?rfc needLines="8" ?>
<section title="SACM Information Model" anchor="sacm-information-model">
<t>The information model defined by the SACM Working Group captures
the types of objects and data required to fulfill the defined
SACM Requirements <xref target="I-D.ietf-sacm-requirements"/>. It additionally
provides details on the flow of data to and from the different
objects in the system, in conjunction with the SACM Architecture
document <xref target="I-D.ietf-sacm-architecture"/>. The document describes
all of these things in a protocol and data format neutral
manner.</t>
<t>The document provides descriptions of the various components that
are required to perform endpoint assessments, along with some
usage scenarios, and the potential mapping from OVAL to any of
these defined components wherever OVAL may be relevant.</t>
</section>
<section title="OVAL Language" anchor="oval-language">
<t>The OVAL Language is made up of several parts, each responsible
for encapsulating a part of the assessment model. Each part is
discussed briefly below <xref target="STRUCTURE-OF-OVAL"/>.</t>
<t>Note: A word about Core vs. Platform Extensions. OVAL can be
broadly split into Core structures, which are those that are
foundational and give the overall structure to the OVAL
Language, and the Platform Extensions, which are
platform-specific structures that extend the Core in order to
provide ways to encode the underlying low-level,
platform-specific tests used by OVAL Content. This paper is
chiefly focused on mapping the Core into the SACM Information
Model.</t>
<t>In a similar fashion, while thinking about how to implement the
SACM Information Model, two distinct levels must be considered:
<list style="numbers">
<t>Platform-agnostic, high level concepts</t>
<t>Platform-specific concepts</t>
</list>
</t>
<section title="Core OVAL Data Models"
anchor="core-oval-data-models">
<t>There are a number of models defined as part of OVAL. This
section discusses the three most important models.</t>
<section title="OVAL Definitions Model"
anchor="oval-definitions-model">
<t>The Definitions Model is the central component of the
OVAL Language. The structures in this model allow an
author to encode what data to collect, the expected
values for the data, and the rules by which to evaluate
that data. This represents one of the chief limitations
of the OVAL Language: Since an author must include both
what data must be collected, and how that data is to be
assessed in an OVAL Definition, these two related-but
separate-elements are directly coupled to one another.
For more information, see the related "Collection and
Evaluation Must Be De-coupled" section below.</t>
<t>It is in the OVAL Definitions Model that the Definition
object is defined. A Definition is the root element for
any OVAL check. It contains a set of criteria, either
simple or complex, to define how the check should
operate. In addition, the OVAL Definitions Model defines
the base structures that are used by the Platform
Extensions to extend OVAL, as well as Functions, and
other high-level concepts.</t>
</section>
<section title="OVAL System Characteristics Model"
anchor="oval-system-characteristics-model">
<t>The OVAL System Characteristics Model defines structures
to encode the actual data that is collected. It provides
basic structures for capturing this data, including the
Item, which is the base structure for recording
collected data in OVAL. It also provides structures for
capturing information about the endpoint from which the
data was collected, including OS information, endpoint
identification information (such as IP and MAC
addresses), and other relevant endpoint metadata.</t>
</section>
<section title="OVAL Results Model" anchor="oval-results-model">
<t>Finally, OVAL provides a third model to encode the
results of the assessment: the OVAL Results Model. In
addition to providing structures to capture essential
information about the evaluation results, such as the
overall results of each assessment and when the
assessment occurred, the model also provides ways to
include both the guidance (Definitions) and collected
data (System Characteristics) used for the assessments.
This model provides a comprehensive way to capture not
only the results themselves, but also the information
used to determine the result.</t>
</section>
</section>
<section title="Additional Core OVAL Data Models"
anchor="additional-core-oval-data-models">
<t>Additional data models are defined that support specific
capabilities and are sometimes useful in conjunction with
the OVAL models previously discussed. The models discussed
in this section are not intended to stand alone, and require
the use of one or more of the core OVAL models.</t>
<section title="OVAL Common Model" anchor="oval-common-model">
<t>The Common Model is a very simple collection of global
building blocks, such as enumerations used throughout
the other models, along with some other foundational
pieces. Common values are defined in this model once and
then applied within other OVAL models, thus reducing
redundancy between each OVAL data model. Examples of the
elements provided by the OVAL Common Model are
enumerations that provide useful value sets for use
within OVAL, such as family types ("windows", "unix",
etc.), data types (e.g., "string," "boolean," "int,"
etc.), and class types (e.g., "vulnerability,"
"compliance," etc.).</t>
</section>
<section title="OVAL Variables Model"
anchor="oval-variables-model">
<t>The OVAL Variables Model provides a simple framework for
externally specifying variable values to an OVAL
Definitions document at runtime.</t>
</section>
<section title="OVAL Directives Model"
anchor="oval-directives-model">
<t>The OVAL Directives Model provides a very simple model
with structures to indicate the level of detail that
should be present in an OVAL Results document.</t>
</section>
</section>
</section>
<section title="Relating the OVAL Models to the SACM Information Model"
anchor="relating-the-oval-models-to-the-sacm-information-model">
<t>The following section discusses each piece of the SACM
Information Model, where one or more OVAL models can be used to
implement the piece, wholly or in part.</t>
<section title="Attribute Collector" anchor="attribute-collector">
<t>The SACM Information Model defines both Internal and External
Attribute Collectors. Both are components that perform the
collection of posture information about an endpoint. The
Information Model lists a number of examples of Collectors
such as Network Intrusion Detection Systems (NIDS), NEA
posture collectors, and vulnerability scanners. While OVAL
is not directly applicable for some types of Attribute
Collectors such as NIDS, it is certainly applicable for
items such as configuration or vulnerability scanners.</t>
<t>An Attribute Collector needs to be instructed as to what
specific posture attributes must be collected, when or how
often those attributes must be collected, and how to share
the collected attributes. In some cases, an Attribute
Collector may simply collect data and directly respond to
the caller with the required results. In others, it may need
to execute the data collection at a future time or at some
interval, and thus will need to know how to share the
collected data. The OVAL Language does not provide any
mechanism for instructing tools where to send collected
data, but the OVAL Definitions Model can (among other
things) encode what data must be collected; however, it does
not allow (as currently constructed) for providing any
notion of what constitutes valid data collection (i.e., how
recent data must be to be considered acceptable, and how and
where it was collected).</t>
<t>Additionally, the OVAL Definitions Model could be modified to
support monitoring of events. As it is today, OVAL doesn't
have any explicit way to include this instruction, but it
would be simple to modify the model to include this
notion.</t>
<t>The OVAL System Characteristics Model allows the encoding of
collected information and can be used to implement a data
format for sharing collected data. While OVAL does not
require that tools store data using a standardized format
(though they are free to do so), a standardized format is
required to allow tools to share data. The OVAL System
Characteristics Model provides a standardized way to encode
this information.</t>
</section>
<section title="Evaluator" anchor="evaluator">
<t>An Evaluator is the component that analyzes inputs such as
Posture Attributes and Evaluation Guidance to determine the
result of a particular assessment. It is the piece that
answers a question about the security posture of one more
endpoints. The Evaluator must be able to ingest inputs of
various types, understand the question or questions asked of
it, and analyze the inputs to make a determination.</t>
<t>In this case, OVAL could be used to provide several of the
required inputs to an Evaluator. The format defined in the
OVAL Definitions Model could be used to express Evaluation
Guidance. Note that when mapping the OVAL Definitions Data
Model to the SACM Information Model, it is important to
distinguish between Collection and Evaluation within the
OVAL Definitions Model. The OVAL Definitions Model
structures currently combine both the Collection ("what to
collect") and Evaluation ("what the data should look like").
One of the key concepts within the SACM Information Model is
that Collection and Evaluation should be separate concepts.
Nonetheless, OVAL contains building blocks that could be
modified to satisfy this need.</t>
<t>Additionally, the structures defined in the OVAL System
Characteristics Model could be used to encode the Posture
Attributes input to the Evaluator. Finally, when the
Evaluator makes its determination and must encode the result
of the assessment, the OVAL Results Model structures could
be used as well.</t>
</section>
<section title="Endpoint Attribute Assertion"
anchor="endpoint-attribute-assertion">
<t>According to the SACM Information Model, an Endpoint
Attribute Assertion is a way to indicate that a specified
set of posture attributes or events were present on an
endpoint during a specific interval of time. For example, an
Assertion could be made that a particular Windows server had
the following attributes from 1/1/2015 - 1/8/2015: <list
style="symbols">
<t>os = Windows 7</t>
<t>mac-address = 01:24:42:58:34:2b</t>
</list>
</t>
<t>OVAL does not have a direct corollary to this construct;
however, the structures defined by the OVAL System
Characteristics Model could provide a base from which such a
construct could be built. The System Characteristics Data
Model is designed to capture posture attributes, and as
such, could be extended or modified to include the concept
of a time interval.</t>
<t>Additionally, it is important to note that the SACM
Information Model also states that Events can be included
within an Endpoint Attribute Assertion. While "event" and
"attribute" are often used interchangeably, in the SACM
Information Model, these two concepts are considered
distinct. The distinction is that an "event" is something
that has a value that does not change until something causes
a change, whereas an "attribute" is something that occurs at
a moment in time. The Endpoint Attribute Assertion deals
with both posture attributes and events during a time
interval. No special treatment is given to Events within
OVAL as it is currently constructed, although, as stated
previously, adding a time interval to support Events is
simple to do.</t>
</section>
<section title="Evaluation Result" anchor="evaluation-result">
<t>An Evaluation Result is the representation of the analysis of
a given set of Posture Attributes against Evaluation
Guidance. The OVAL Results Model structures can be used to
encode one or more Evaluation Results.</t>
</section>
<section title="Collection Guidance" anchor="collection-guidance">
<t>Within the SACM Information Model, Collection Guidance is
defined as information that describes which Posture
Attributes must be collected from one or more endpoints. It
is the means by which an Attribute Collector determines what
information it must collect, as well as when that
information must be collected (including intervals for
repeated collection activities).</t>
<t>The OVAL Definitions Model provides structures capable of
expressing information about what data must be collected for
an assessment. It is important to note that the method by
which the OVAL Definitions Model accomplishes this will not
necessarily directly apply to the SACM Information Model in
its current state. In many cases, which specific posture
attributes should be collected is not distinct from its
evaluation guidance. For the OVAL Definitions Model to be
used to implement the SACM Information Model, work would
need to be undertaken to de-couple these concepts.</t>
<t>While the model provides the ability to encode details such
as what data must be collected from the endpoint, it does
not currently provide the ability to include information
such as collection interval. The model can be extended,
however, to add this capability. Adding the concept of an
"interval" to the model to capture the concept may be a way
to accomplish this goal.</t>
<t>Important Note: One of the key drawbacks to OVAL is that
Platform Extensions (using the OVAL Definitions Model as a
base) must be created for each platform and data source to
capture any Posture Attributes that must be collected for a
given platform and data source. As a result, it is not easy
or scalable to create or update extensions for rapidly
changing platforms and products in a timely manner.</t>
<t>With this in mind, it is important that any use of the OVAL
Definitions Model to satisfy Collection Guidance for SACM
should warrant consideration of updates that change this
from a solution where the low-level platform details are
part of the language itself, to one where the format
provides a way for domain experts (ideally primary source
vendors) to instruct tools what Posture Attributes to
collect.</t>
<t>This also applies to the next section (Evaluation
Guidance).</t>
</section>
<section title="Evaluation Guidance" anchor="evaluation-guidance">
<t>The Evaluation Guidance component contains the information
that directs an Evaluator how to perform one or more
assessments based on collected data. Evaluation Guidance
must direct the Evaluator on what the expected state of
collected data should be. Additionally, it must be able to
specify desired characteristics of the data. That is, it
must be able to not only cite the specific posture
attributes under evaluation, but also to specify
characteristics such as the type of tool that was used to
collect the data, how old the data is, etc.</t>
<t>The Evaluator must then ingest this guidance, locate the
required data-whether locally or remotely available-and then
execute the analysis required.</t>
<t>OVAL offers the OVAL Definitions Model to provide the
structures for encoding the expected state or values for the
collected data. The OVAL Language does not currently provide
a way to specify the expected characteristics of the data,
but the OVAL Definitions Model could be augmented to include
this type of information. Alternatively, the concept could
be added elsewhere and re-used as appropriate. Allowing for
characteristics information will be important to allow
evaluation to do things like only query data if it's been
collected within the past x days or only query data that is
collected by a credentialed scan.</t>
<t>Again, as Collection and Evaluation are intertwined currently
in OVAL Language, some work will be required to de-couple
them for use with the Evaluation Guidance component.</t>
</section>
<section title="Report Generator" anchor="report-generator">
<t>Within the SACM Information Model, a Report Generator is a
component that constructs a collection of artifacts such as
Endpoint Attribute Assertions, Evaluation Results, etc. The
reports that are generated by this component can be used to
either report on a collection of assessments, or to provide
a summary of previously run assessments or queries.</t>
<t>The structures defined in the OVAL Results Model could be
used by the Report Generator as a means to encode the
required report information. The OVAL Results structures can
be used to encode Evaluation Results, although it would need
to be extended to include Endpoint Attribute Assertions; as
such, a construct does not exist within OVAL today.</t>
</section>
<section title="Report" anchor="report">
<t>A Report is the tangible output from the Report Generator. It
contains formatted information that satisfies one or more
assessments or queries. A data format that implements the
Report component must be able to support Evaluation Results,
Endpoint Attribute Assertions, Events, and other related
reporting data.</t>
<t>As mentioned in the Report Generator section, the OVAL
Results Model structures can be used to encode some of these
details today, and can be extended to handle Events and
Endpoint Attribute Assertions. It is important to note that
while OVAL provides this capability, historically, the OVAL
Results Model structures have been considered by many to be
too verbose to be used in a large scale enterprise
environment. These structures could be revised to allow more
granular and flexible ways to encode this information in
order to satisfy the needs of SACM, or a different data
model could be used to implement this capability.</t>
</section>
<section title="Provenance" anchor="provenance">
<t>Within the SACM Information Model, Provenance is defined as a
metadata item that contains information such as the time
when an artifact is produced, what produced it, the policies
that govern it, and the method used to produce it.</t>
<t>Within the OVAL Common Model, a Generator structure is
defined to express both what created the content, and when
it was created. While the purpose of this structure does not
meet all the needs for Provenance in SACM, it could be used
as a building block to achieve this goal. This structure
would need to be extended to include the policy information
and the method used to produce a given artifact, but the
base structure is in place.</t>
</section>
</section>
<section title="SACM Constructs with No OVAL Mapping"
anchor="sacm-constructs-with-no-oval-mapping">
<t>Finally, while there are many similarities between what is
defined by the SACM Information Model and that supported by OVAL
models, there are some things discussed in the SACM Information
Model document that are either different from-or not supported
within-OVAL.</t>
<section title="Tasking" anchor="tasking">
<t>The SACM Information Model discusses Tasks in a few places,
including the Collector, Evaluator, and Reporting sections.
Tasks represent of notion of "do something at this time."
OVAL does not support any notion of a tasking model as
currently defined.</t>
<t>While the OVAL Definitions Model (or some derivative) could
be referenced by a model that captures tasking, it may be
difficult to support all of the needs of tasking in this
way. Tasking may already be well defined by another,
existing model, and if so, it might be best to leverage that
existing work.</t>
</section>
<section title="Event-driven Actions" anchor="event-driven-actions">
<t>Within the SACM Information Model, in addition to posture
attributes, events are also often part of the data
collection activities. Events are discussed as both part of
an Endpoint Attribute Assertion, and an Endpoint Attribute
Collector. In each case, it is clear that, in addition to
the collection of posture attribute data, event data must
also be taken into account.</t>
<t>The OVAL Language does not have any notion of capturing
events directly. It is constructed to allow the
representation of Posture Attribute data within the OVAL
System Characteristics Model, but event data is absent from
that model. OVAL can be modified to support Events in large
part by simply extending it to include a time interval.</t>
</section>
<section title="User and Authorization"
anchor="user-and-authorization">
<t>The Information Model talks about Users (i.e., one or more
end users or roles) and Authorizations (i.e., their
authority to undertake actions). While OVAL includes some
entities that may relate to these types of concepts, they
appear in very specific low-level tests like Windows and
UNIX user-related tests. OVAL lacks any general concept of
Users or Authorizations that could be applied across its
core data structures. The recommendation is to integrate an
external solution into relevant OVAL models to achieve
required capabilities in this area.</t>
</section>
<section title="Location" anchor="location">
<t>Similar to Users and Authorization, Locations are defined in
the Information Model. Locations include authentication
points, wall-jacks to which an endpoint is connected,
geographical location, etc.</t>
<t>Again, as for Users and Authorization, the recommendation is
for the relevant OVAL models to be integrated with other
solutions to meet these requirements.</t>
</section>
</section>
<section title="Lessons Learned and Gaps"
anchor="lessons-learned-and-gaps">
<t>Over the course of ten-plus years in moderating the OVAL project,
those involved in the project have released over 15 distinct
versions of the Language, 25 versions of the OVAL Interpreter,
and have processed over 25,000 OVAL Definitions in the OVAL
Repository. In addition, the team has spent a lot of time
interacting with security tool vendors, researchers, primary
source vendors, and commercial and government end users,
discussing their needs and struggles. As such, the following
lessons learned are presented to help ensure that the collective
experience of the group is shared with the larger community.</t>
<t>In addition to a description of the lesson, each also has a
suggested application for the SACM work.</t>
<section title="Simplicity is Key" anchor="simplicity-is-key">
<section title="Lesson" anchor="simplicity-is-key-lesson">
<t>Endpoint assessment covers a broad set of activities.
From organization to organization, assessment has
different meanings, and what is "good enough" for one
group, barely scratches the surface for another.
Experience suggested that caution must be used to avoid
unnecessary complexity as a means to address this
diversity.</t>
<t>The team has seen that when information sharing is
required across diverse parties, the simpler the
exchange mechanism design, the more successful the
sharing effort will be.</t>
</section>
<section title="SACM Implications"
anchor="simplicity-is-key-sacm-implications">
<t>Review both the diversity of the different organizations
that are sharing information within the SACM framework,
and the types and volume of information that must be
shared. Include only the information that is required to
successfully implement the desired Use Cases.</t>
</section>
</section>
<section title="Collection and Evaluation Must Be De-coupled"
anchor="collection-and-evaluation-must-be-de-coupled">
<section title="Lesson"
anchor="collection-and-evaluation-must-be-de-coupled-lesson">
<t>As OVAL - and the security automation space in general -
has evolved, it has become clear that the close coupling
found in OVAL between the OVAL Object and OVAL State
(i.e., what to collect and what the collected data
should look like) is an undesirable feature. By forcing
these two concepts into a single model, the Language
does not easily allow for dynamic querying of previously
collected data, nor does it easily allow for
efficiencies in data collection.</t>
</section>
<section title="SACM Implications"
anchor="collection-and-evaluation-must-be-de-coupled-sacm-implications">
<t>Keep the mechanism by which data is collected and
evaluated separate.</t>
</section>
</section>
<section title="Keep Separate Core and Extensions"
anchor="keep-separate-core-and-extensions">
<section title="Lesson"
anchor="keep-separate-core-and-extensions-lesson">
<t>OVAL, by design, must be frequently updated to keep up
with new and expanding sets of assessment platforms.
However, tool vendors incurred great cost in updating to
new versions of the Language, including implementing new
tests in the updated version, as well as general quality
testing, updating release and deployment, etc.</t>
<t>As the project matured, so too did the Core Models that
define the building blocks for endpoint assessment. Over
the past few years, the Core Models rarely changed-in
some cases, going years without any required update. The
Platform Extension Models, however, will always require
a frequent revision cycle, and often were out of date
very quickly. Despite the fact that these two models had
distinct release cycle requirements-one continually
getting longer in the Core Models, and one requiring
agility in the Platform Extensions-a full release of
both was required to include changes to any part of the
OVAL Language.</t>
</section>
<section title="SACM Implications"
anchor="keep-separate-core-and-extensions-sacm-implications">
<t>SACM should focus on providing the foundational building
blocks that allow those that know best to express what
data must be collected to assess an endpoint. The SNMP
standard <xref target="RFC1157"/> could be used as a model for this
type of separation. SNMP defines the building blocks for
sharing information about network devices, but defers
the low-level details of this information sharing to
those that best understand the products via Management
Information Bases (MIBs). While this is not a perfectly
analogous model for the SACM work, this clean separation
of core building blocks and protocols from the low-level
details of products should be emulated, if possible.</t>
</section>
</section>
<section title="Empower Subject Matter Experts"
anchor="empower-subject-matter-experts">
<section title="Lesson"
anchor="empower-subject-matter-experts-lesson">
<t>As the security automation field has matured, more
primary source vendors and other subject matter experts
have taken increased responsibility in ownership of how
their products are assessed. This step in maturity is
critical and, within OVAL, as these vendors have become
more involved, the quality in tests available to tools
and end users has greatly increased.</t>
</section>
<section title="SACM Implications"
anchor="empower-subject-matter-experts-sacm-implications">
<t>Ensure that usage of SACM means that those that best
understand the component being assessed are empowered to
instruct what data must be collected for the assessment,
along with the meaning of this data. As much as
possible, keep the mechanism by which this information
is conveyed as simple as possible to ensure that it is
as easy as possible for subject matter experts to
participate.</t>
</section>
</section>
<section title="Carrots Work Better than Sticks"
anchor="carrots-work-better-than-sticks">
<section title="Lesson"
anchor="carrots-work-better-than-sticks-lesson">
<t>As much as possible, ensure that usage and compliance
with the defined standards is encouraged by offering
primary source vendors and subject matter experts
incentive to do so. Forced compliance typically
encourages organizations to do the least possible, and
does not entice them to continually stay engaged.</t>
</section>
<section title="SACM Implications"
anchor="carrots-work-better-than-sticks-sacm-implications">
<t>Find ways to encourage participation that drives long
term engagement and willing participation. Engage with
vendors to understand their problems and, where
possible, construct SACM use cases and requirements that
not only address the needs of the SACM end users, but
also those of the vendors. Build a compelling story for
use of SACM that not only shows value to end users, but
shows a clear return on investment for vendors.</t>
</section>
</section>
<section title="Use Caution Defining Data Collection"
anchor="use-caution-defining-data-collection">
<section title="Lesson"
anchor="use-caution-defining-data-collection-lesson">
<t>When providing information about what data must be
collected as part of an assessment, it can be quite easy
to provide this information in a way that dictates how
to collect the required data. Doing so can limit
innovation and architectural choices for organizations
implementing security automation tools.</t>
<t>On the other hand, it is not always feasible to express
what data must be collected without implying or
instructing specific data collection mechanisms. Over
the years, there have been a few cases where the OVAL
community could not agree on significant issues related
to data collection. Discussions on whether to allow open
scripting in the Language and how best to support both
third party and primary source contributions were very
challenging. With good arguments on both sides of these
issues, it was difficult to achieve consensus.</t>
</section>
<section title="SACM Implications"
anchor="use-caution-defining-data-collection-sacm-implications">
<t>This will be one of the bigger challenges for SACM to
navigate. SACM must allow those that best understand
platforms and products to instruct what data must be
collected for assessment. At the same time, third party
support will be critical in some cases as well, and
allowances must be made for this.</t>
<t>Additionally, deciding how many, if any, collection
methods are allowed as part of the collection
instructions will be challenging. Again, a balance
should be struck to best allow clarity in data
collection instructions, without limiting innovation and
product-specific decisions.</t>
</section>
</section>
<section title="Perspective Matters" anchor="perspective-matters">
<section title="Lesson" anchor="perspective-matters-lesson">
<t>When evaluating collected posture attributes, it is
important to be able to include additional context to
this evaluation in some cases. For example, the method
by which data was collected could be an important piece
of information when performing evaluation. If the
scanner was a remote, unauthorized scanner of an
endpoint, it is entirely possible that the scanner could
not properly scan for a number of posture attributes.
If, however, the scanner ran locally on the endpoint as
an administrative user, it is much more likely that it
accurately collected posture attributes from the
endpoint.</t>
<t>Other examples of this type of perspective and context
include how old the collected data is, and whether the
scanner was active or passive.</t>
</section>
<section title="SACM Implications"
anchor="perspective-matters-sacm-implications">
<t>Ensure information that provides necessary context can be
provided as part of data collection, thereby allowing
context-based decisions to be made.</t>
</section>
</section>
<section title="Flexible Reporting Fidelity is Important"
anchor="flexible-reporting-fidelity-is-important">
<section title="Lesson"
anchor="flexible-reporting-fidelity-is-important-lesson">
<t>After data collection and evaluation is complete, the
results of the evaluation must be shared, often with
multiple parties, and in multiple ways. It is important
to provide a reasonable amount of flexibility with
respect to what levels of fidelity are allowed with
results. While OVAL did try to achieve a reasonable
amount of flexibility with reporting fidelity,
challenges still exist.</t>
</section>
<section title="SACM Implications"
anchor="flexible-reporting-fidelity-is-important-sacm-implications">
<t>As much as possible, allow the end users of the reporting
capability to determine exactly what level of fidelity
they need to achieve their goals.</t>
</section>
</section>
<section title="Evaluation Guidance is Platform-Specific"
anchor="evaluation-guidance-is-platform-specific">
<section title="Lesson"
anchor="evaluation-guidance-is-platform-specific-lesson">
<t>In the early days of OVAL, initial adoption of the effort
was spearheaded by third party security vendors, as
opposed to the primary source vendors for software. As
the effort matured, more primary source vendors became
involved and adopted OVAL in some way. It quickly became
evident that, while third party vendors made great
strides in determining how to evaluate the security
posture of many platforms and products, understanding
the best way to evaluate is hard, and very
platform-specific. Additionally, OVAL content is costly
to create, even for seasoned content authors, due to the
need to understand these very low-level product and
platform complexities.</t>
</section>
<section title="SACM Implications"
anchor="evaluation-guidance-is-platform-specific-sacm-implications">
<t>As cited above, the primary source vendors are best
suited to provide evaluation guidance. It is very
challenging for third party organizations to truly
understand platform-specific evaluation. Empower primary
source vendors and other subject matter experts by
providing simple and effective ways to provide this
information. Also, as discussions on complex topics
arise, engage these primary source vendors to understand
their valuable views.</t>
</section>
</section>
</section>
<section title="Recommendations" anchor="recommendations">
<t>In order to successfully standardize the mechanisms by which
endpoint posture assessment is performed, the following
recommendations are offered to SACM for consideration.</t>
<section
title="Use the OVAL System Characteristics Model for Encoding Collection Data"
anchor="use-the-oval-system-characteristics-model-for-encoding-collection-data">
<t>The OVAL System Characteristics Model is used within the OVAL
Language in order to encode the underlying data collected as
part of endpoint posture assessment. Each of the posture
attributes collected by an OVAL-enabled tool can be
represented using the OVAL System Characteristics Model. As
such, this model should be used as the basis for
implementing at least one of the formats used to encode
collected posture attributes within SACM.</t>
<t>Within the OVAL System Characteristics Model, information
such as metadata about the document (who/what created the
document, creation timestamp, etc.), endpoint identification
information (OS name, host name, and other asset-related
information), and the foundational constructs to allow the
encoding of posture attributes can be found. It is
understood that modifications to the model will be required
in order for it to fully implement all of the requirements
for SACM. However, the use of this well-supported,
standardized mechanism for encoding collected data is
recommended as SACM begins moving from Information Model
into Data Models and actual implementations.</t>
<t>The expectation is that SACM will need to make use of
multiple types of standardized formats to encompass a
complete solution for endpoint posture assessment. As such,
the OVAL System Characteristics Model is likely to be used
as one of multiple possible formats for encoding collected
data.</t>
</section>
<section
title="Use the OVAL Definitions Model for Collection and Evaluation Guidance"
anchor="use-the-oval-definitions-model-for-collection-and-evaluation-guidance">
<t>Similar to the OVAL System Characteristics Model, the OVAL
Definitions Model also has aspects that could be very useful
in jump starting the development of a model to capture
Collection Guidance. Collection Guidance is the mechanism by
which a content author can dictate what rules should be used
for collecting data from an endpoint. While the OVAL
Definitions Model, as it is today, is used for guidance of
both Collection and Evaluation, it is well suited to serve
as the base for Collection Guidance.</t>
<t>This model provides several key features that should be used
as building blocks for this capability. For instance, within
the OVAL Definitions Model, there is a series of structures
that can serve as the base for instructing tools as to what
data must be collected, including abstract structures for
identifying required posture attributes, Variables, and
Functions (which allow several types of data manipulation
during collection). The model also supports a number of
different data types, such as strings, Booleans, integers,
records, and others.</t>
<t>While the recommendation is to make use of many of the
structures found within the OVAL Definitions Model, it is
equally important to note that the current approach for
extending OVAL into various platforms is flawed, and should
be fixed. Specifically, for every new check that is to be
added to the Language, a new concrete test must be created.
OVAL provides an abstract Test structure that must be
extended to create checks (e.g., "registry_test,"
"file_test," "ldap_test," etc.). For SACM, it is imperative
that a more scalable and flexible approach be
implemented.</t>
<t>One aspect of SACM that has been discussed, but only
partially worked into the Information Model at the time, is
the concept of high-level, platform-agnostic configuration
items and low-level platform-specific configuration items.
In the discussed concept, the high-level items will capture
the concepts of configuration that must be defined by those
who write the guidance, while the low-level items will be
provided by the appropriate vendors and/or subject matter
experts to allow those that best know the platforms and
products to instruct data collection. With this approach in
place, some of the concepts defined within the OVAL
Definitions Model (e.g., Objects, which instruct tools as to
what data to collect) will need to be modified or removed to
accommodate the shift in how posture attributes are defined
for Collection. As such, the recommendation is to use many
of the underlying structures in the OVAL Definitions Model,
including the data types, Variables, Functions, etc., as a
base from which to build a complete solution for fulfilling
the SACM Information Model.</t>
<t>In addition to utility in supporting Collection Guidance, the
same OVAL Definitions Model should also be used as the base
for Evaluation Guidance. Again, with the current OVAL
Language, Collection and Evaluation are wrapped together in
the single model. The OVAL Definitions Model provides a
series of structures that can be used to support Boolean
logic statements, which could be useful for defining
evaluation criteria and could be used as the basis for a
further enhanced model for Evaluation.</t>
</section>
<section
title="Do NOT Use the OVAL Results Model for Results Sharing"
anchor="do-not-use-the-oval-results-model-for-results-sharing">
<t>Despite the fact that the Results Model could be used to
share the results of the evaluation part of an endpoint
posture assessment, the recommendation is to not use this
model to represent this information within SACM. The OVAL
Results Model has, over the years, been a source of
contention at times within the OVAL Community. Some feel
like it provides too little information, while others
believe that it offers too much. While there is some
flexibility, in the form of OVAL Directives, in how much or
how little information is included in the results, it really
is not flexible enough to handle the broad set of
requirements for SACM without extensive re-working.</t>
<t>Furthermore, SACM is working hard at separating data
collection and evaluation, which makes the OVAL Results
Model a poor fit, as it was constructed with a more combined
Collection and Evaluation framework. It is expected that to
properly model all of the results requirements within SACM,
an alternative solution will be required.</t>
<t>While considering an alternative way to encode the results of
an assessment, the following requirements have been stated
by the OVAL Community as critical factors: <list
style="symbols">
<t>Allow evaluation results with appropriate
granularity</t>
<t>Ensure support for enterprise scale uses</t>
<t>Provide results that include only the actionable
information</t>
<t>Ensure that data is clear and identifiable within the
results</t>
<t>Ensure interoperability</t>
</list>
</t>
</section>
</section>
<section anchor="Acknowledgements" title="Acknowledgements">
<t>The authors would like to thank Brant Cheikes (MITRE), Juan
Gonzalez (DHS), Adam Montville (CIS), Charles Schmidt (MITRE),
David Waltermire (NIST), and Kim Watson (DHS) for reviewing this
document and providing helpful feedback.</t>
</section>
<!-- Possibly a 'Contributors' section ... -->
<section anchor="IANA" title="IANA Considerations">
<t>This memo includes no request to IANA.</t>
</section>
<section anchor="Security" title="Security Considerations">
<t>This memo documents, for informational purposes, the mapping
between the OVAL Data Models and the SACM Information Model as
well as the lessons learned from the past 10+ years of
developing OVAL. As a result, there are no specific security
considerations.</t>
</section>
<section anchor="ChangeLog" title="Change Log">
<section anchor="latest" title="-00 to -01">
<t>There are no textual changes associated with this revision.
This revision simply reflects a resubmission of the document so that
it goes back into active status. The document expired on
November 6, 2015.</t>
</section>
</section>
</middle>
<!-- *****BACK MATTER ***** -->
<back>
<!-- References split into informative and normative -->
<!-- There are 2 ways to insert reference entries from the citation libraries:
1. define an ENTITY at the top, and use "ampersand character"RFC2629; here (as shown)
2. simply use a PI "less than character"?rfc include="reference.RFC.2119.xml"?> here
(for I-Ds: include="reference.I-D.narten-iana-considerations-rfc2434bis.xml")
Both are cited textually in the same manner: by using xref elements.
If you use the PI option, xml2rfc will, by default, try to find included files in the same
directory as the including file. You can also define the XML_LIBRARY environment variable
with a value containing a set of directories to search. These can be either in the local
filing system or remote ones accessed by http (http://domain/dir/... ).-->
<references title="Normative References">
<!--?rfc include="http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml"?-->
&RFC2119; </references>
<references title="Informative References">
<!-- Here we use entities that we defined at the beginning. -->
<!-- A reference written by by an organization not a person. -->
<reference anchor="SACM"
target="https://datatracker.ietf.org/wg/sacm/charter/">
<front>
<title>IETF Security Automation and Continuous Monitoring
(sacm) Working Group Charter</title>
<author>
<organization>The IETF SACM WG</organization>
</author>
<date year="2015"/>
</front>
</reference>
<reference anchor="SACM-DOCUMENTS"
target="https://datatracker.ietf.org/wg/sacm/documents/">
<front>
<title>IETF Security Automation and Continuous Monitoring
(sacm) Working Group Documents</title>
<author>
<organization>The IETF SACM WG</organization>
</author>
<date year="2015"/>
</front>
</reference>
<reference anchor="OVAL-WEBSITE" target="https://oval.mitre.org/">
<front>
<title>The Open Vulnerability and Assessment
Language</title>
<author>
<organization>The MITRE Corporation</organization>
</author>
<date year="2015"/>
</front>
</reference>
<reference anchor="OVAL-DEFINITION-TUTORIAL"
target="http://oval.mitre.org/language/about/definition.html">
<front>
<title>The OVAL Definition Tutorial</title>
<author>
<organization>The MITRE Corporation</organization>
</author>
<date year="2011"/>
</front>
</reference>
<reference anchor="I-D.ietf-sacm-requirements"
target="http://www.ietf.org/id/draft-ietf-sacm-requirements-04.txt">
<front>
<title>Secure Automation and Continuous Monitoring (SACM)
Requirements</title>
<author>
<organization>Cam-Winget, N. and L.
Lorenzin</organization>
</author>
<date year="2015"/>
</front>
</reference>
<reference anchor="I-D.ietf-sacm-architecture"
target="http://www.ietf.org/id/draft-ietf-sacm-architecture-03.txt">
<front>
<title>Secure Automation and Continuous Monitoring (SACM)
Architecture</title>
<author>
<organization>Cam-Winget, N., Ford, B., Lorenzin, L.,
McDonald, I., and l. loxx@cisco.com</organization>
</author>
<date year="2015"/>
</front>
</reference>
<reference anchor="STRUCTURE-OF-OVAL"
target="http://oval.mitre.org/language/about/structure.html">
<front>
<title>Structure of the Language</title>
<author>
<organization>The MITRE Corporation</organization>
</author>
<date year="2012"/>
</front>
</reference>
<reference anchor="RFC1157"
target="https://www.ietf.org/rfc/rfc1157.txt">
<front>
<title>A Simple Network Management Protocol (SNMP)</title>
<author>
<organization>Case, J., Fedor, M., Schoffstall, M., and
J. Davin</organization>
</author>
<date year="1990"/>
</front>
</reference>
</references>
<!-- Change Log
v00 2006-03-15 EBD Initial version
v01 2006-04-03 EBD Moved PI location back to position 1 -
v3.1 of XMLmind is better with them at this location.
v02 2007-03-07 AH removed extraneous nested_list attribute,
other minor corrections
v03 2007-03-09 EBD Added comments on null IANA sections and fixed heading capitalization.
Modified comments around figure to reflect non-implementation of
figure indent control. Put in reference using anchor="DOMINATION".
Fixed up the date specification comments to reflect current truth.
v04 2007-03-09 AH Major changes: shortened discussion of PIs,
added discussion of rfc include.
v05 2007-03-10 EBD Added preamble to C program example to tell about ABNF and alternative
images. Removed meta-characters from comments (causes
problems).
2015-04-17 AR updated ipr attribute. -->
</back>
</rfc>
| PAFTECH AB 2003-2026 | 2026-04-24 07:30:24 |