Operating System
Operating System
Table of contents
1 Protection Profile Introduction..........................................................................................................9
1.1 Protection Profile reference.....................................................................................................9
1.2 TOE overview..........................................................................................................................9
1.2.1 TOE type......................................................................................................................10
1.2.2 Hardware / software / firmware supporting the TOE...................................................10
1.3 Structure of the Protection Profile.........................................................................................11
1.4 Terminology...........................................................................................................................12
1.4.1 Users.............................................................................................................................12
1.4.2 Groups..........................................................................................................................12
1.4.3 Subjects........................................................................................................................13
1.4.4 Resources.....................................................................................................................14
1.4.5 Objects.........................................................................................................................14
1.4.6 Security attributes........................................................................................................15
1.4.7 Trusted users / subjects................................................................................................15
1.4.8 Security policy.............................................................................................................16
1.4.9 Storage object types.....................................................................................................16
1.5 References..............................................................................................................................16
2 OSPP Framework............................................................................................................................17
2.1 Mandatory information given by the ST................................................................................17
2.1.1 Conformance claim......................................................................................................17
2.1.2 SFR reference with OSPP extended package reference...............................................17
2.2 Mandatory information given by OSPP extended packages..................................................18
2.2.1 Extended package identification..................................................................................18
2.2.2 Extended package composition rules...........................................................................18
2.2.3 Specification of OSPP extended packages...................................................................18
2.3 Specification restricted to the OSPP base..............................................................................18
3 OSPP Base Introduction..................................................................................................................20
3.1 TOE overview........................................................................................................................20
3.1.1 Auditing........................................................................................................................20
3.1.2 Cryptographic services.................................................................................................21
3.1.3 User data protection.....................................................................................................22
3.1.4 Identification and authentication..................................................................................23
7.1.6 Rationale......................................................................................................................42
7.2 FDP_RIP.3 Full residual information protection of resources...............................................42
7.2.1 Component leveling.....................................................................................................42
7.2.2 Management.................................................................................................................42
7.2.3 Audit.............................................................................................................................43
7.2.4 FDP_RIP.3 Full residual information protection of resources.....................................43
7.2.5 Rationale......................................................................................................................43
7.3 FIA_USB.2 Enhanced user-subject binding..........................................................................43
7.3.1 Component leveling.....................................................................................................43
7.3.2 Management.................................................................................................................43
7.3.3 Audit.............................................................................................................................43
7.3.4 FIA_USB.2 Enhanced user-subject binding................................................................43
7.3.5 Rationale......................................................................................................................44
8 Security Requirements.....................................................................................................................45
8.1 Security functionality provided by remote trusted IT systems..............................................45
8.1.1 Example: Access control policy...................................................................................46
8.2 Security Functional Requirements.........................................................................................48
8.2.1 FAU_GEN.1 Audit data generation.............................................................................48
8.2.2 FAU_GEN.2 User identity association........................................................................50
8.2.3 FAU_SAR.1 Audit review...........................................................................................50
8.2.4 FAU_SAR.2 Restricted audit review...........................................................................50
8.2.5 FAU_SEL.1 Selective audit.........................................................................................50
8.2.6 FAU_STG.1 Protected audit trail storage....................................................................51
8.2.7 FAU_STG.3 Action in case of possible audit data loss...............................................51
8.2.8 FAU_STG.4 Prevention of audit data loss...................................................................51
8.2.9 FCS_CKM.1(SYM) Cryptographic key generation....................................................52
8.2.10 FCS_CKM.1(RSA) Cryptographic key generation...................................................53
8.2.11 FCS_CKM.1(DSA) Cryptographic key generation...................................................53
8.2.12 FCS_CKM.2(NET) Cryptographic key distribution..................................................54
8.2.13 FCS_CKM.4 Cryptographic key destruction.............................................................54
8.2.14 FCS_COP.1(NET) Cryptographic operation..............................................................54
8.2.15 FDP_ACC.1(PSO) Subset access control..................................................................55
8.2.16 FDP_ACC.1(TSO) Subset access control..................................................................56
8.2.17 FDP_ACF.1(PSO) Security attribute based access control........................................56
Index of Tables
Table 1: Coverage of security objectives for the TOE.......................................................................36
Table 2: Coverage of security objectives for the TOE environment..................................................37
Table 3: TOE threats sufficiency........................................................................................................38
Table 4: Security policies sufficiency.................................................................................................39
Table 5: Assumptions sufficiency.......................................................................................................41
Table 6: Security Functional Requirements coverage........................................................................73
Table 7: Security Functional Requirements rationale.........................................................................75
Table 8: Security Functional Requirements dependency analysis......................................................78
Illustration Index
Illustration 1: Types of TOE instances and their boundaries..............................................................29
Revision History
Version Date Author Changes
1.9 2010-04-06 Stephan Müller, atsec Initial public release
2.0 2010-06-01 Andreas Siegert, atsec Changed to support accessibility requirements
The main purpose of a general-purpose operating system (from a security point of view) is to
provide defined objects, resources and services to entities using the functions provided by the
operating system at its external interfaces, and to enforce a defined policy on access to objects, use
of resources, and use of services. At a minimum, the operating systems addressed by this Protection
Profile export interfaces to programs executing "on top of” the operating systems and interfaces to
external entities, including network interfaces, as well as interfaces to devices that are used to
"transport" data or actions of external entities to the operating system (for example, a keyboard and
a mouse). In addition, the operating system uses functions of the underlying hardware and software
to provide its functions, including using devices that are not connected to an external entity such
that this entity could affect the behavior of the device directly (for example, hard disks or displays).
An operating system conformant to this Protection Profile can be operated as a server system within
a data center, but also as a client system used directly by one or more human users. While it is
mandatory that an operating system conformant to this Protection Profile must be capable of
providing and using some basic network services, such a system may also be started in an
environment where it is not connected to any network and with the network services inactive. It is
mandatory that an operating system conformant to this Protection Profile must provide basic
security functionality for user identification and authentication, access control, management and
audit.
The TOE will provide user services directly or serve as a platform for networked applications, and
will support protected communication using one or more cryptographically-protected network
protocols or the support of dedicated, physically-separated network links. To support protected
communication, the TOE must implement at least the TCP/IP network protocol family; this
Protection Profile makes no statements about the version of IP.
The OSPP addresses general-purpose operating systems operating in a well-managed enterprise
environment. This addresses mostly servers, but also desktop clients if their operating environment
fulfills the security problems defined in chapter 4, as well as the security problems defined by any
OSPP extended packages claimed in the ST. These security problems include requirements for
professional management of the system and basic protection against physical attacks that can be
found in enterprise or government environments, but typically not in home environments
administered by private users. The enterprise or government environments may include setups for
mobile systems or home-offices provided that the TOE implements mechanisms that allow these
environments to comply with the security problem definition in this PP. The OSPP makes no claims
or statements that it specifically applies to either a server operating system or a client operating
system. If an operating system meets the requirements defined in the security problem definition of
the OSPP base, with or without any extended packages, the operating system can claim
conformance to this Protection Profile.
virtualization layer. Such virtualization emulates all or part of the hardware in a manner that is
either transparent to the TOE or by having the TOE using dedicated interfaces to the virtualization
layer. In any case, the interfaces to the underlying platform must be defined and described to allow
analysis of how the operating system uses the functionality of the underlying platform.
At a minimum, the underlying platform must provide functions the operating system can use to
protect itself from untrusted subjects interfering with the functionality of the operating system or
bypassing its protection functions. This requires functions that allow the operating system to:
⚫ Protect areas of main memory from being accessed by untrusted subjects.
⚫ Protect devices from being directly accessed (without that access being mediated by the
operating system) by untrusted subjects.
⚫ Protect any other function of the underlying platform from being used by untrusted subjects in a
way that would violate the security policy of the operating system.
This Protection Profile does not define how the underlying platform implements those mandatory
protection functions.
At a minimum, the TOE boundary encompasses all parts of the operating system software that are
capable of bypassing all or parts of the claimed protection functions. Many operating systems are
structured into a “kernel” operating with privileges of the underlying hardware to configure
memory, processor states and devices; and a set of "trusted subjects" that operate with privileges
assigned by the kernel that allow those trusted subjects to violate all or parts of the security policy
the whole operating system needs to enforce. Such trusted subjects also must be considered as part
of the TSF.
The TSF subject to assessment may be augmented with OSPP extended packages adding useful
security functionality.
In the view of this Protection Profile, the underlying platform is located in the IT environment. This
does not preclude a conformant ST from drawing the TOE boundary differently by including all or
parts of the underlying platform. For example, an ST author may decide to include the virtualization
layer into the TOE, but still exclude the underlying hardware.
1.4 Terminology
The following sections define terminology for the Operating System Protection Profile (OSPP).
1.4.1 Users
As defined in the Common Criteria, users are external entities that interact with the TOE. Such
external entities include human users, as well as other IT systems.
Users can be either anonymous (that is, the operating system does not know the identity of the user)
or they may be associated with an identity. In all cases where the security policy enforced by the
operating system distinguishes between different users, the operating system must be sure that the
identity of the user is correct.
It is quite common that an operating system supports different types of users. Those different types
of users are allowed to use different sets of interfaces, have different security attributes, are
identified and authenticated in different ways, and are subject to different rules of the security
policy. For example, an IT system as a "user" may only be allowed to connect via defined network
services, is authenticated using a challenge-response protocol that makes use of digital certificates,
and is not allowed to directly access file system objects. On the other hand, "human users" are
allowed to use the system call interfaces (via subjects bound to them), are authenticated using a
userid/password combination (and eventually some other authentication mechanisms), and are
allowed to directly access (via a subject started on behalf of the user) file system objects in
accordance with the rules of a discretionary access control policy for those objects.
Users may be locally defined and managed. In this case, the operating system must maintain a list
of valid users with their security attributes and must have a policy that defines how those users are
managed.
In many cases, an operating system also allows users that are not locally-defined and managed to
connect to the operating system and request services. In those cases, the operating system relies on
another trusted IT system to ensure the following:
⚫ The user is still a valid member of the user community and has not been revoked.
⚫ User security attributes passed to the operating system by a remote trusted entity are still valid.
Note that user security attributes may be passed to the TOE within a digital certificate. In this
case, the certification authority that issued the digital certificate is the remote trusted entity,
even though the TOE may never have a direct connection to this entity.
1.4.2 Groups
Groups define a set of users that can be referred to by a group identifier. Like users, groups may be
managed either by the TOE itself or by a remote trusted entity. Management of groups includes:
1.4.3 Subjects
Subjects are the active entities in the system. With regard to the execution of programs, an OSPP-
conformant operating system must allow for identifying and separating different active entities
executing "on top of” the operating system into different "subjects" that are uniquely identifiable by
the operating system, allowing the operating system to control the subject's access to objects,
allocation of resources, and use of operating system services by enforcing the rules of a defined
policy. The architecture of an OSPP-conformant operating system must prevent such subjects from
violating any of the policy rules or bypassing the controls within the operating system that enforce
the policy rules.
The operating system may recognize "trusted subjects" for which some or all of the policy rules are
not enforced. Such "trusted subjects", when part of the evaluated configuration, must be part of the
TSF. Such "trusted subjects" must not provide a way for untrusted users to violate the rules of the
security policy.
This Protection Profile does not prescribe how an operating system implements, separates and
controls the subjects it creates. This aspect must be explained in the Security Target and then further
elaborated in the evidence presented for the security architecture assurance component.
An operating system conformant to this Protection Profile must be able to "bind" specific external
entities ("users") to the subject. Subjects bound to a user are operating on his behalf. Since the
policy rules enforced by the operating system are often defined by "user security attributes", the
operating system must have rules that define how the security attributes of a subject operating on
behalf of a user are derived when the operating system "binds" the subject to the user. In the
simplest case, the user security attributes are copied one-to-one to the subject security attributes.
Significantly more complex rules are implemented in many operating systems. For example, an
operating system may have rules that define how the subject security attributes are derived from the
user security attributes, the security attributes of the active groups the user is a member of, as well
as the environment in which the subject is started (which may include the time and date or the port
the user has used to connect to the TOE), and the current state of the TOE. This Protection Profile
does not prescribe the rules for user-subject binding. Therefore, those rules must be defined in a
Security Target that claims conformance to this Protection Profile.
Note that operating systems themselves may create and use subjects that are actively involved in the
enforcement of the security policy or that are able to bypass all or part of the policy. These subjects
need to be "trusted" to enforce the defined policy and are, therefore, part of the TSF of the operating
system. In addition, some operating systems create subjects that are part of the TSF upon creation,
but change to "untrusted" subjects afterwards (for example, as part of the process of binding a user
to the subject).
Subjects may be created by the operating system that are not bound to any user, for example,
daemons that are started by the operating system either during start-up or as a result of specific
events. For these subjects, the operating system must have a policy that defines the active set of
privileges and access rights for these subjects in order to be able to consistently enforce the rules of
the security policy. Some operating systems use a mechanism of "pseudo-users", whereby subjects
are started with the identity of a "user" without this identity being assigned to any real user. This
allows the operating system to use the functions of user management to assign privileges and access
rights and to use the rules for user-subject binding to establish the active set of privileges and access
rights for these subjects. Since pseudo-users do not represent external entities, usually no user
authentication is required.
1.4.4 Resources
Resources are a finite set of logical and/or physical entities that the operating system may allocate
to users, subjects or objects. Resource allocation must be managed by the TOE. Blocks of persistent
storage, CPU cycles, main memory, and network bandwidth are examples of resources. Resources
are usually allocated and if they are re-usable, later de-allocated and prepared for re-use. The OSPP
base does not require a specific policy covering resources to be implemented and how they are
allocated to subjects, users or objects. However, the OSPP base requires that all re-usable resources,
when allocated to a different subject, user or object than the one it was last allocated to, must be
prepared for re-use such that upon re-allocation, no information can be obtained from the resource
about its previous use or content. OSPP extended packages may define more restricted resource
clearing mechanisms, such as the clearing of the contents of a resource upon de-allocation. OSPP
extended packages may also require the implementation of specific policies for allocating resources,
for example, management of quotas or specific priorities when allocating resources.
1.4.5 Objects
Objects are passive entities created and controlled by the operating system, which provide services
to users and/or subjects to use those objects. Named objects are covered by the operating system
implementing an access control policy enforcing rules that define the conditions that must be met
for users and/or subjects to use a specific type of named objects in a defined way. Named objects
must have an identifier that allows the operating system to identify the object when a subject
attempts to access the object or when the security attributes of or access rights to the object are
managed. Please note that objects may exist or be instantiated by the TOE without being accessible
to subjects. For such TOE-internal objects, the security policy of the TOE may not apply as long as
they remain internal objects.
The OSPP base requires that at least one type of named object must be created and maintained in
persistent storage and must allow users and/or subjects to:
⚫ Create a new object of this type
⚫ Write data to an object
⚫ Read data from an object
⚫ Delete an object
Other operations on this type of named object may be defined, but are not mandatory in the OSPP
base.
For this type of named object, the OSPP base requires that an access control policy must be
implemented that clearly defines the conditions that must be met to allow a user and/or subject to
perform one of the four defined operations on an object of this type. Further conditions the access
control policy must meet are defined later in this document.
An operating system usually implements a number of different types of named objects and may
implement a different access control policy for each named object type.
In addition to trusted users, an operating system may also have trusted subjects. Similar to trusted
users, these are subjects that have the capability to bypass some or all of the rules defined in the
security policy or the capability to manage TSF data on which the security policy relies. These
subjects may either not be bound to a user, or they may be bound to a user and allow this user to
access objects and/or resources he is not allowed to access when bound to an untrusted subject.
Trusted subjects, therefore, have additional capabilities that untrusted subjects do not have, and they
enforce a subject-specific policy on the use of such capabilities. An example is a trusted subject that
allows a user to modify specific TSF data (for example, his own password). Because of their
additional capabilities, trusted subjects are part of the TSF.
1.5 References
The following references are applicable to this document, as well as all OSPP extended package
documents unless a reference is re-defined.
CC Common Criteria for Information Technology Security Evaluation
Parts 1 through 3, July 2009, Version 3.1 Revision 3
CEM Common Methodology for Information Technology Security
Evaluation, July 2009, Version 3.1 Revision 3
2 OSPP Framework
The OSPP allows the definition of functional extensions that can be optionally claimed by an ST in
addition to the OSPP base. As such, the OSPP defines the following components:
⚫ The OSPP base specifies the conformance claim, security problem, objectives, and security
functional requirements that are to be implemented by every general-purpose operating system.
The OSPP base is mandatory and defines the common denominator for all operating systems
claiming conformance with the OSPP.
⚫ An OSPP extended package specifies the security problem definition, objectives, and security
functional requirements for mechanisms that may be implemented in addition to the OSPP base.
Usually, an OSPP extended package defines an extension that is either desired or implemented
by several general-purpose operating systems. However, the functionality specified in an OSPP
extended package is not commonly found among general-purpose operating systems. OSPP
extended packages can optionally be added to the OSPP base functionality when writing an ST.
The ST author may choose from the set of OSPP extended packages when deriving an ST. To
avoid fragmentation of security functionality into OSPP extended packages that are too small to
be practical, an OSPP extended package shall define a set of functional requirements that
address one or more general security problems.
The OSPP is defined as an extensible framework. The current set of OSPP extended packages can
be enhanced with newly-developed or updated OSPP extended packages. Those will then be part of
a re-evaluation and re-certification of the OSPP base. Therefore, this framework invites anybody
interested in specifying an aspect of general-purpose operating systems to author an OSPP extended
package and commit it to the OSPP forum, where the OSPP is managed. Using this approach, there
will always be a valid set of OSPP base and extended packages, which are compliant to each other.
Dependencies on other OSPP extended packages can be specified.
3.1.1 Auditing
All operating systems conformant with this Protection Profile must implement audit functionality
that allows the operating system to record events viewed as security-relevant. The records created
by the operating system for such events must contain at least the type of the event, the time the
event occurred, the identity of the user or subject that caused the event (where appropriate), and
further event-specific data. If the event is a request to use a function, the record also needs to
contain sufficient information about how the function was intended to be used (usually defined by
the parameter passed to the function) and the outcome of the function. If the event is related to an
operation performed on an object, the identity of the object must be contained in the record.
Audit records must be stored in an audit trail in persistent storage unless they are transmitted to a
trusted centralized audit server, as indicated below. Local storage used for the audit trail must be
protected from unauthorized access by users or subjects. A policy must exist that defines:
⚫ The actual events to be audited (from the overall list of auditable events)
⚫ Rules that define when a user or subject can define the events to be audited
⚫ Rules that define when a user or subject can read audit records from the audit trail
⚫ Rules that define when a user or subject can delete or re-initialize the audit trail
The operating system must monitor the amount of space allocated to the audit trail and take
appropriate actions when it detects that it has insufficient space to store further audit records.
The audit generation functionality is completely provided locally (by the TSF exclusively). The
TOE shall be able to:
⚫ Gather audit information from security-relevant events
⚫ Provide functionality to store audit information locally, and potentially provide a remote storage
mechanism (analysis of audit data applies to locally-stored audit data only)
⚫ Provide local analysis of the audit trail if the trail is stored locally
⚫ Allow selection of which audit records are to be generated
⚫ Provide protection of the audit trail when stored locally
⚫ Provide protection that no audit records are lost
Note that remote audit handling is moved to an OSPP extended package. In addition, a TOE can use
remote functions to store and/or evaluate audit data and allow appropriately authorized users to
define which of the different audit capabilities are used.
control policies may be applicable to the same portion of persistent storage. If the operating system
does not resolve such conflicts automatically, the guidance must explain how to set appropriate
access rights such that the two access control policies do not conflict.
The OSPP requires the ST author to specify the default access rights for new subjects, as well as
new access-controlled objects.
Finally, the OSPP requires the ST author to specify the rules the TOE enforces before allowing a
user or subject to manage TSF data used within the access control rules. Usually those rules are
based on specific TSF data (like user privileges). If this TSF data can be managed, the management
rules that apply also must be specified. It is up to the ST author to describe the conditions that must
be satisfied in order to manage TSF data (including the TSF data used in the access control rules).
The OSPP allows locally- and remotely-stored TSF data to be used within the access control rules.
In addition, the OSPP allows the ST author to specify whether the TOE provides access control
decisions for other remote trusted IT products. With this option, the ST author can specify the server
side of permission storage.
and authentication, an operating system will perform a user-to-subject binding whenever it starts an
untrusted subject that shall operate on behalf of the authenticated user.
The OSPP requires that a user or another IT system must be authenticated before utilizing any
services of the operating system that are restricted by the security policy to specific users. An
OSPP-conformant system may allow unauthenticated users to access objects controlled by the
access control policy. This requires the access control policy to be able to assign access rights to
unauthenticated users.
An operating system may accept users as identified and authenticated when another trusted IT
system reports the identity of the user in a way that allows the operating system to verify the
integrity and authenticity of the message that containing the information about the remotely
authenticated user.
An operating system may also authenticate users with the help of another trusted IT system, for
example, when it either retrieves information used for the authentication from the other system (for
example, the hash value of a password), or redirects information it retrieves from the user to the
other system such that the remote trusted IT system can perform the user authentication and report
the result back to the TOE.
For the OSPP base, the TOE shall provide identification and authentication services by allowing
locally- and remotely-performed identification and authentication with the following definitions:
⚫ Local identification and authentication implies that the TOE performs the operations to
establish the identity of the user. This definition allows storing the TSF data holding the user's
credentials either on the TOE or on a remote trusted IT system. However, the TOE must be able
to completely fetch the TSF data with the credential information and perform the necessary
operations and checks that implement the identification and authentication logic locally.
Another local identification and authentication is performed when a user provides a token (a
certificate, Kerberos token, etc.) which defines the user's identity; the TOE must verify that
token.
⚫ Remote identification and authentication implies that the TOE is a client to an authentication
server. The TOE sends the user-supplied identification and authentication data to the server and
queries the server as to whether the transmitted credentials are positively or negatively verified.
The TOE then enforces the decision made by the authentication server.
⚫ For accessing public objects, the TOE shall allow operations by unauthenticated users (which
shall be exempt from identification and authentication). The allowed operations and public
objects must be defined by the ST author.
For example, a Directory Server may store the user credentials or the internal representations of the
user credentials. When the TOE is able to obtain all credentials, including the user password, and
performs the operations to validate the user-given credentials with the stored ones, then a local
identification and authentication is performed. However, if the TOE only performs, for example, an
LDAP-bind operation with the user-supplied credentials and observes whether the LDAP server
rejects the operation, then remote identification and authentication is performed.
The OSPP allows for local, remote and combined local and remote identification and authentication,
which can usually be found in large installations. For example, a local user database is defined with
administrative user IDs that are only usable when the connection to the authentication server is
severed. Another example would be that the TOE caches the user database of the authentication
server and applies this database in case the link to the authentication server is severed. Note that if
the TOE allows multiple authentication methods concurrently (such as local and remote
authentication), the ST author shall specify the order in which the authentication methods are
applied.
In addition, the OSPP shall allow the ST author to specify whether the TOE provides identification
and authentication for other remote trusted IT products. With this option, the ST author can specify
the server side of the credential storage.
When credentials or the internal representations of the user credentials are stored within the TOE,
the TOE shall ensure the quality of the credentials when they are being changed by administrative
users or authorized users.
At a minimum, the identification and authentication functionality shall provide all of the following
mechanisms:
⚫ User ID / password
⚫ Software token-based authentication
After successful identification and authentication, the TSF may perform a user-subject binding.
Such a binding is required when the operating system creates and starts a subject to operate on
behalf of the user. This process ensures that the external entity (or user) "binds" to the subject. The
ST author must define the rules applicable to the user-subject binding process. Those rules define
how the security attributes of the subject are initialized, usually derived from security attributes of
the user. See section 1.4.3 for more details. During the user-subject binding, all security attributes of
a subject used by the rules of the security policy must be established.
⚫ Management of security attributes that are used for the information flow control policy.
⚫ Management of the audit policy, which includes at least the selection of the events to be audited
and the management of the storage objects that contain the audit trail.
The OSPP does not mandate any specific implementation. However, the TOE must:
⚫ Allow administrative functions to be assigned to zero, one, or more users.
The ST author shall specify the rules used to determine if a management activity is allowed and the
TSF data used in those rules. The OSPP does not specify any policy or any specific set of rules. As
such, the ST author has the ability to specify one user that is granted all privileges (like the UNIX
root user). In addition, the ST author can also specify a sophisticated administration policy
including hierarchical privileges or role-based management.
The TOE shall allow localized and/or centralized management of these security functions:
⚫ Localized management implies that tools are provided with the TOE to configure aspects of
security functions. The SFRs will not make any statement about whether the TOE data is stored
remotely (see discussion about local identification and authentication above).
⚫ Remote management implies that the management of the security functionality is not provided
by the TOE, but the TOE enforces the management actions.
Irrespective of the management type (localized or remote), the configuration data can be stored
locally or remotely. If stored remotely, there is no restriction about whether the configuration data is
stored with the remote management system or on another system.
The OSPP allows locally- and remotely-stored TSF data used within the management rules.
In addition, the OSPP allows the ST author to specify whether the TOE provides management
decisions for other remote trusted IT products. With this option, the ST author can specify the server
side of the management operations.
Such co-operation between trusted IT systems is not necessarily restricted to a client-server type of
relationship. It may also be a peer-to-peer relationship, for example, where a smartcard is used as
part of the user authentication process. In addition, an operating system may make use of multiple
remote trusted IT systems to provide a single security functionality: to authenticate a user, the TOE
may require the user to present a smartcard. The smart card (as the representative of the user) may
present a digital certificate, in response to which the TOE may use a challenge-response protocol to
verify that the smart card actually contains the private key associated with the digital certificate it
presents. Furthermore, the TOE may use the services of a Directory Server to validate that the
certificate has not been revoked. In addition, the TOE may also use its peer-to-peer connection to a
cryptographic module outside of the TOE boundary in order to perform the cryptographic
operations required for the smart card authentication process (including the validation of the digital
signature of the CA that issued the certificate presented by the smart card) and the process to
validate the digital signature of the certificate revocation list provided by the Directory Server.
External Entity External Entity External Entity External Entity External Entity
(human or (human or (human or (human or (human or
system) system) system) system) system)
4 Conformance Claims
The following sections describe the conformance claims of the Operating System Protection Profile
(OSPP).
5.1 Threats
Threats to be countered by the TOE are characterized by the combination of an asset being subject
to a threat, a threat agent and an adverse action.
The definition of threat agents and protected assets that follows is applicable to the OSPP base, as
well as to the OSPP extended packages, unless noted otherwise.
5.1.1 Assets
Assets to be protected are:
⚫ Persistent storage objects used to store user data and/or TSF data, where this data needs to be
protected from any of the following operations:
⚪ Unauthorized read access
⚪ Unauthorized modification
⚪ Unauthorized deletion of the object
⚪ Unauthorized creation of new objects
⚪ Unauthorized management of object attributes
⚫ Transient storage objects, including network data
⚫ TSF functions and associated TSF data
⚫ The resources managed by the TSF that are used to store the above-mentioned objects,
including the metadata needed to manage these objects.
The TOE protects against intentional and unintentional breach of TOE security by attackers
possessing an enhanced-basic attack potential.
The following threats are addressed by the OSPP base-conformant TOEs. The PP covers these
threats and organizational security policies necessary to derive the necessary security functionality.
There are no threats and policies to justify the assurance level. This is deemed unnecessary, since
the chosen evaluation assurance level is already defined in the CC with a rationale explaining the
threats countered by the assurance measures.
5.3 Assumptions
The specific conditions below are assumed to exist in a PP-conformant TOE environment.
6 Security Objectives
The following sections describe the security objectives of the Operating System Protection Profile.
O.TRUSTED_CHANNEL The TSF must be designed and implemented in a manner that allows
for establishing a trusted channel between the TOE and a remote
trusted IT system that protects the user data and TSF data transferred
over this channel from disclosure and undetected modification and
prevents masquerading of the remote trusted IT system.
T.IA.USER The threat of accessing user data, TSF data or TOE resources
without being identified and authenticated is removed by:
⚫ O.I&A requiring that each entity interacting with the TOE is
properly identified and authenticated before allowing any
action the TOE has defined to provide to authenticated users
only.
P.USER The policy to match the trust given to a user and the actions the
user is given authority to perform is implemented by:
⚫ O.MANAGE allowing appropriately-authorized users to
manage the TSF,
⚫ OE.INFO_PROTECT, which requires that users are trusted to
use the protection mechanisms of the TOE to protect their data.
7.1.3 Management
There are no management activities foreseen.
7.1.4 Audit
There are no actions defined to be auditable.
7.1.6 Rationale
The quality of the random number generator is defined using this SFR. The quality metric required
in FCS_RNG.1.2 is detailed in the German Scheme AIS 20 and AIS 31.
7.2.2 Management
See management description specified for FDP_RIP.2 in [CC].
7.2.3 Audit
See audit requirement specified for FDP_RIP.2 in [CC].
7.2.5 Rationale
FDP_RIP.3 addresses the problem of resources implemented in main memory that may be allocated
to and de-allocated from subjects or users. Unless those resources lose their content automatically
as part of the de-allocation and re-allocation process, they must be subject to a process that prepares
them for re-use by rendering the previous content unavailable to the subject or user to which it is
next allocated. An example is main memory that has been allocated to a subject; this memory must
be cleared before it can be re-allocated to a subject with different security attributes (for example a
subject operating on behalf of a different user). This preparation prevents the passing of security-
critical information via this resource, since such unregulated passing would potentially allow the
subject or user to which the memory is next allocated to use this information to violate the security
policy. Typical examples of such critical information that may be passed via resources not prepared
for re-use are passwords or cryptographic keys.
7.3.2 Management
See management description specified for FIA_USB.1 in [CC].
7.3.3 Audit
See audit requirement specified for FIA_USB.1 in [CC].
FIA_USB.2.2 The TSF shall enforce the following rules on the initial association of user
security attributes with subjects acting on the behalf of users: [assignment:
rules for the initial association of attributes].
FIA_USB.2.3 The TSF shall enforce the following rules governing changes to the user
security attributes associated with subjects acting on the behalf of users:
[assignment: rules for the changing of attributes].
FIA_USB.2.4 The TSF shall enforce the following rules for the assignment of subject
security attributes not derived from user security attributes when a subject is
created: [assignment: rules for the initial association of the subject security
attributes not derived from user security attributes].
7.3.5 Rationale
An operating system may derive subject security attributes from other TSF data that are not directly
user security attributes. An example is the point-of-entry the user has used to establish the
connection. An access control policy may also use this subject security attribute within its access
control policy, allowing access to critical objects only when the user has connected through specific
ports-of-entry.
8 Security Requirements
This chapter specifies the requirements set forth for the TOE. If the OSPP mandates a specific
option that cannot be specified as part of the SFR or SAR, the PP marks it as “ST Author Note”.
The ST author must apply this note when writing an ST and claiming conformance with this PP.
Notes marked as “Application Note” are informative to support the understanding of the SFR or
SAR.
The following styles of marking operations are applied with this Protection Profile:
⚫ Assignments and selections are marked in bold face font.
⚫ Iterations are marked by appending a suffix to the SFR identification.
⚫ Refinements are marked in bold and italic face font.
system and waits for its response. Based on the response, actions are taken. These actions can
be as simple as enforcing the decision (in the example above, the user is allowed to log in or
not), or these actions can be more complex, for example, the TSF might perform additional
processing based on the response.
The OSPP requires the following consideration of the such security mechanisms that rely on remote
trusted IT systems:
⚫ If the TSF implements the client aspect, the ST author must specify:
⚪ One or more assumptions outlining the behavior expected from the remote trusted IT
system, specified such that a different ST author (the ST author who may characterize the
server side) can derive SFRs from this specification. Note that the specification shall be as
precise as necessary for the TSF. The ST author might want to specify a precise rule set to
be implemented by the server counterpart, but it might also be possible that the ST author
only specifies a very generic reply type the TSF expects from the server.
⚪ The SFR specification must only cover the locally visible (at the TSFI) policy without any
hint to the policy enforced by the remote trusted IT system. Note that the specification of
this security policy is in addition to any security policies mandated by the OSPP base
(extended packages may be specified requiring a pre-defined client-side security policy).
This implies that such a security policy covering the client side of a general security policy
must extend the OSPP base. It is not permissible that such a client side security policy is
used to cover SFRs mandated by the OSPP base, as the client side functionality must be
provided in addition to the general-purpose computing environment mandated by the OSPP
base.
⚫ If the TSF implements the server aspect, the ST author must specify:
⚪ SFRs implementing the security policy that comply with the assumptions specified in the
ST for the client side. Note that the specification of this security policy is in addition to any
security policies mandated by the OSPP base as outlined for the client-side above.
⚪ The TSF data of the client side covering the respective security functionality usually
transform into user data on the server side, because for the server side, the transmitted data
are not used to implement or enforce its security policy. For such scenarios, it is beneficial
to define extended components on the server side that resemble the respective client-side
SFR.
⚫ The client side implements the access control enforcement by controlling and knowing the
subjects, objects, and operations between subjects and objects. When an operation is requested
from a subject on an object that is covered by the access control policy, the client side submits a
corresponding request to the remote trusted IT system providing the above-mentioned
information and waits for a response. Once the response is received, the client side enforces the
response by either allowing or denying the requested operation.
The example must be covered in the ST specifying the TOE operating as a client:
A.REMOTE_PSO A remote trusted IT system implements and provides access control
decisions to the TSF for the following security policy:
Access control policy decisions are returned to the TSF with the TSF
providing the parameters with each request consisting of [subject
security attribute of user ID, object security attribute of file system
object name, and operation requested by the subject on the object].
FDP_ACC.2.1 The TSF shall enforce the Extended Persistent Storage Object Access
Control Policy on
a) Subjects: processes acting on behalf of users;
b) Objects: Persistent Storage Objects of file system object;
and all operations among subjects and objects covered by the SFP.
FDP_ACC.2.2 The TSF shall ensure that all operations between any subject controlled by the
TSF and any object controlled by the TSF are covered by an access control
SFP.
FDP_ACF.1.1 The TSF shall enforce the Extended Persistent Storage Object Access
Control Policy to objects based on the following:
a) Subject security attributes: user ID;
b) Persistent storage object attributes: file system object name.
FDP_ACF.1.2 The TSF shall enforce the following rules to determine if an operation among
controlled subjects and controlled objects is allowed:
a) Access to the object by the subject is granted if the remote trusted IT
system that is provided with the parameters of [subject security
attribute, object security attribute, and operation requested by the
subject on the object] replies that the operation is to be allowed.
FDP_ACF.1.3 The TSF shall explicitly authorize access of subjects to objects based on the
following additional rules: none.
FDP_ACF.1.4 The TSF shall explicitly deny access of subjects to objects based on the
following additional rules: If the remote trusted IT system is not available
or does not respond within 30 seconds after the access request is
submitted, access is denied.
The example must be covered in the ST specifying the TOE operating as a server using the
following SFRs defined as extended components (the extension is due to the removal of the
dependency on FDP_ACC.1, as the attributes are not assigned to subjects and objects covered by
the TSF):
FDP_ACF_EXT.1.1 The TSF shall support the Extended Persistent Storage Object Access
Control Policy for requesting clients based on the following:
a) Subject security attributes:
i. user ID transmitted by a remote trusted IT system;
ii. group ID assigned to the user ID;
b) Persistent storage object attributes:
i. file system object name transmitted by a remote trusted IT
system;
ii. file system object owning user ID;
iii. file system object owning group ID;
iv. file system object permissions.
FDP_ACF_EXT.1.2 The TSF shall apply the following rules to determine if an operation among
controlled subjects and controlled objects is allowed:
a) Per-user access granting: The TSF returns to the requester the
indication that access to the object by the subject is granted if the
subject security attribute of user ID is equal to the object security
attribute of the file system object owning user ID and the
permissions assigned to the subject attribute – object attribute
combination allows the operation requested by the subject for the
owning user; or
b) Per-group access granting: The TSF returns to the requester the
indication that access to the object by the subject is granted if the
subject security attribute group ID is equal to the object security
attribute of the file system object owning group ID and the
permissions assigned to the subject attribute – object attribute
combination allows the operation requested by the subject for the
owning group.
FDP_ACF_EXT.1.3 The TSF shall explicitly authorize access of subjects to objects based on the
following additional rules: none.
FDP_ACF_EXT.1.4 The TSF shall explicitly deny access of subjects to objects based on the
following additional rules: none.
1
[selection, choose one of: minimum, basic, detailed, not specified]
2
[assignment: other specifically defined auditable events].
3
[assignment: other audit relevant information]
Application Note: The subject identity may be identical to the user identity in the case where the
subject identity is established by the user-subject binding process. In this case,
only one identity needs to be included in the audit record. The purpose here is
the ability to trace an event to the user that caused the event. This may not be
possible if the subject identity does not allow to identify the user the subject
was bound to when the event happened. In order to support FAU_GEN.2, the
user identity has, therefore, been added as the information to be recorded.
Application Note: The outcome to be recorded with the audited event can either be binary
(success or failure) or the value resulting from the event, depending on the
implementation of the TOE. For example, access control decision shall store
the information about the result of the access control decision with the audit
trail. A TOE may implement more decision results than just access allowed or
denied, where all of these results shall be recorded as outcome of the access
control check event.
4
[assignment: authorised users]
5
[assignment: list of additional attributes that audit selectivity is based upon]
6
[selection: object identity, user identity, subject identity, host identity, event type]
7
[assignment: list of additional attributes that audit selectivity is based upon]
sends a notification to the remote trusted IT systems sending audit data to the
TOE to inform about this state.
15
[assignment: cryptographic key sizes]
16
[assignment: list of standards]
17
[assignment: list of standards]
18
[assignment: cryptographic key destruction method]
19
[assignment: list of standards]
Application Note: The ST author shall consult the national scheme for acceptable key destruction
standards.
b) Objects:
i. Persistent Storage Objects of the following type [assignment:
enumerate all persistent storage objects covered by this SFP];
ii. [assignment: other storage objects covered by this SFP];
c) Operations: [assignment: list of operations covered by the SFP]23.
ST Author Note: The list of operations on the object needs to cover the creation of a new object,
the destruction of an object, all types of access to the object, as well as
operations on TSF data associated and stored with the object (for example,
object name, access control list associated with the object, other object security
attributes). If some of those operations are covered by SFRs related to the
management of TSF data, the ST author shall include a reference to those SFRs
in order to allow the ST reader to identify where those operations are
described.
ST Author Note: When the TOE includes several different types of named objects implemented
using persistent storage and when the TOE implements fundamentally different
access control rules for those different types of named objects, the ST author
shall use iterations of FDP_ACC to describe the access control rules for the
different object types.
Application Note: A persistent storage object establishes a data storage or data exchange link
between two or more subjects. Examples of persistent storage objects are: files,
directories.
objects controlled under the indicated SFP, and for each, the SFP-relevant
security attributes, or named groups of SFP-relevant security attributes].
FDP_ACF.1.2 The TSF shall enforce the following rules to determine if an operation among
controlled subjects and controlled objects is allowed:[assignment: rules
governing access among controlled subjects and/or users and controlled
objects using controlled operations on controlled objects that allow to grant
access down to the granularity of single subjects or users].
FDP_ACF.1.3 The TSF shall explicitly authorise access of subjects to objects based on the
following additional rules: [assignment: rules, based on security attributes or
other TSF data, that explicitly authorise access of subjects to objects].
FDP_ACF.1.4 The TSF shall explicitly deny access of subjects to objects based on the
following additional rules: [assignment: rules, based on security attributes or
other TSF data, that explicitly deny access of subjects to objects].
and all operations that cause that information to flow to and from subjects
covered by the SFP.
FDP_IFC.2.2 The TSF shall ensure that all operations that cause any information in the TOE
to flow to and from any subject in the TOE are covered by an information flow
control SFP.
Application Note: The OSPP explicitly does not specify the version of the Internet Protocol. This
implies that the Internet Protocol versions usable in the evaluated configuration
must be covered by this SFR.
32
[assignment: additional information flow control SFP rules]
33
[assignment: access control SFP(s) and/or information flow control SFP(s)]
FDP_ITC.2.4 The TSF shall ensure that interpretation of the security attributes of the
imported user data is as intended by the source of the user data.
FDP_ITC.2.5 The TSF shall enforce the following rules when importing user data controlled
under the SFP from outside the TOE: [assignment: additional importation
control rules].
Application Note: If, for example, file names or file name extensions are used for access control
decisions, they are security attributes. In the case that an external file system is
mounted, this is considered an import of user data with security attributes, and
therefore, FDP_ITC rules must be defined and satisfied.
Application Note: Based on the wording of FDP_ITC.2.1, the TOE complies with this SFR even
when it does not allow import of objects covered by the persistent or transient
storage object control policy.
However, the network information flow control policy must always be covered
by the TOE, as it applies to the networking capability of the TOE to control
traffic originating from outside the TOE. In this case, the interpretation of
security attributes is defined by the respective protocol family.
34
[selection: [assignment: positive integer number], an administrator configurable positive integer within
[assignment: range of acceptable values]]
38
[assignment: list of TSF mediated actions]
39
[assignment: list of multiple authentication mechanisms]
40
[assignment: rules describing how the multiple authentication mechanisms provide authentication]
Application Note: For the term “software token verification data”, see the application note for
FIA_ATD.1(HU).
Application Note: To support the specification of the SFR of FIA_UAU.5, the ST author may
want to specify the link mechanism between the frontends and the backends for
providing the authentication mechanisms. For example, the ST author may
wish to state that a login application uses GSSAPI to utilize Kerberos.
41
[assignment: list of feedback]
42
[assignment: list of user security attributes]
FIA_USB.2.3 The TSF shall enforce the following rules governing changes to the user
security attributes associated with subjects acting on the behalf of users:
[assignment: rules for the changing of attributes].
FIA_USB.2.4 The TSF shall enforce the following rules for the assignment of subject
security attributes not derived from user security attributes when a subject is
created: [assignment: rules for the initial association of the subject security
attributes not derived from user security attributes].
43
[assignment: access control SFP(s), information flow control SFP(s)]
44
[selection: change_default, query, modify, delete, [assignment: other operations]]
45
[assignment: list of security attributes]
46
[assignment: the authorised identified roles]
47
[assignment: access control SFP(s), information flow control SFP(s)]
48
[selection: change_default, query, modify, delete, [assignment: other operations]]
49
[assignment: list of security attributes]
50
[assignment: the authorised identified roles]
51
[assignment: access control SFP, information flow control SFP]
52
[selection, choose one of: restrictive, permissive, [assignment: other property]]
53
[assignment: the authorised identified roles]
54
[assignment: access control SFP, information flow control SFP]
55
[selection, choose one of: restrictive, permissive, [assignment: other property]]
FMT_MSA.3.2 The TSF shall allow the [assignment: the authorised identified roles, or
users that satisfy the following rules: [assignment: rules that define when a
user is allowed to override the default values]]56 to specify alternative initial
values to override the default values when an object or information is created.
56
[assignment: the authorised identified roles]
57
[assignment: access control SFP, information flow control SFP]
58
[assignment: the authorised identified roles]
59
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
60
[assignment: list of TSF data]
61
[assignment: the authorised identified roles]
62
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
63
[assignment: list of TSF data]
64
[assignment: the authorised identified roles]
65
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
66
[assignment: list of TSF data]
67
[assignment: the authorised identified roles]
68
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
69
[assignment: list of TSF data]
70
[assignment: the authorised identified roles]
71
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
72
[assignment: list of TSF data]
73
[assignment: the authorised identified roles]
74
[selection: change_default, query, modify, delete, clear, [assignment: other operations]]
75
[assignment: list of TSF data]
76
[assignment: the authorised identified roles]
90
[assignment: specification of revocation rules]
91
[assignment: list of management functions to be provided by the TSF]
92
[assignment: the authorised identified roles]
and provides assured identification of its end points and protection of the
channel data from modification or and disclosure using the following
mechanisms:
a) Cryptographically-protected communication channel using
[assignment: defined cryptographic protocol];
b) [selection: physically protected communication channel;
c) [assignment: other mechanisms for trusted communication channels]].
ST Author Note: The ST author must ensure that the cryptographic protocol specified in this
SFR is also listed in FCS_COP.1(NET).
FTP_ITC.1.2 The TSF shall permit [selection: the TSF, another trusted IT product] to initiate
communication via the trusted channel.
FTP_ITC.1.3 The TSF shall initiate communication via the trusted channel for all security
functions specified in the ST that interact with remote trusted IT systems
and [assignment: list of functions or other conditions which require a
trusted channel]94.
8.3.1.1 Audit
The TOE shall implement a general audit mechanism. This audit mechanism shall generate audit
records for all security-relevant events, where an authorized user shall have the capability to select
the audited events. An authorized user shall be provided with the means to read and interpret the
audit data. The TOE shall protect the audit trail and ensure that proper actions are taken when the
audit trail fills up or is full.
SFR Objectives
FDP_ACC.1(TSO) O.SUBJECT.COM
FDP_ACF.1(PSO) O.DISCRETIONARY.ACCESS
FDP_ACF.1(TSO) O.SUBJECT.COM
FDP_IFC.2(NI) O.NETWORK.FLOW
FDP_IFF.1(NI) O.NETWORK.FLOW
FDP_ITC.2 O.DISCRETIONARY.ACCESS
O.SUBJECT.COM
O.NETWORK.FLOW
FDP_RIP.2 O.AUDITING
O.CRYPTO.NET
O.DISCRETIONARY.ACCESS
O.SUBJECT.COM
O.NETWORK.FLOW
O.I&A
FDP_RIP.3 O.AUDITING
O.CRYPTO.NET
O.DISCRETIONARY.ACCESS
O.SUBJECT.COM
O.NETWORK.FLOW
O.I&A
FIA_AFL.1 O.I&A
FIA_ATD.1(HU) O.I&A
FIA_ATD.1(TU) O.NETWORK.FLOW
FIA_SOS.1 O.I&A
FIA_UAU.1 O.I&A
FIA_UAU.5 O.I&A
FIA_UAU.7 O.I&A
FIA_UID.1 O.I&A
O.NETWORK.FLOW
FIA_USB.2 O.I&A
FMT_MSA.1(PSO) O.MANAGE
FMT_MSA.1(TSO) O.MANAGE
SFR Objectives
FMT_MSA.3(PSO) O.MANAGE
FMT_MSA.3(TSO) O.MANAGE
FMT_MSA.3(NI) O.MANAGE
FMT_MSA.4(PSO) O.MANAGE
FMT_MTD.1(AE) O.MANAGE
FMT_MTD.1(AS) O.MANAGE
FMT_MTD.1(AT) O.MANAGE
FMT_MTD.1(AF) O.MANAGE
FMT_MTD.1(NI) O.MANAGE
FMT_MTD.1(IAT) O.MANAGE
FMT_MTD.1(IAF) O.MANAGE
FMT_MTD.1(IAU) O.MANAGE
FMT_REV.1(OBJ) O.MANAGE
FMT_REV.1(USR) O.MANAGE
FMT_SMF.1 O.MANAGE
FMT_SMR.1 O.MANAGE
FPT_STM.1 O.AUDITING
FPT_TDC.1 O.DISCRETIONARY.ACCESS
O.SUBJECT.COM
O.NETWORK.FLOW
FTA_SSL.1 O.I&A
FTA_SSL.2 O.I&A
FTP_ITC.1 O.TRUSTED_CHANNEL
The dependencies for security assurance requirements are all fulfilled based on the following facts:
⚫ EAL4 is completely self-sufficient with all dependencies being fulfilled with the package of
EAL4.
⚫ The security assurance requirement of ALC_FLR.3 which is in addition to EAL4 does not have
any dependencies.
⚫ The refinement of ASE_CCL.1 does not introduce any dependencies.
Rationale for unresolved dependencies:
⚫ FMT_MSA.3(NI): FMT_MTD.1(NI) is specified to require the management of security
attributes for the Network Information Flow Control Policy, just as a potential
FMT_MSA.1(NI) would have been specified. However, the Network Information Flow Control
Policy is not required to be enforced when managing the security attributes, as the management
aspect of the network information flow control functionality is not protected by the network
information flow control mechanism. Therefore, FMT_MSA.1 is not applicable and is replaced
with FMT_MTD.1(NI).
9 Abbreviations
Abbreviation Description
AH Authentication Header
CC Common Criteria
DAC Discretionary Access Control
EAL Evaluation Assurance Level
ESP Encapsulating Security Payload
IKE Internet Key Exchange
IPSEC IP Security Protocol
MAC Mandatory Access Control
OSPP Operating System Protection Profile
PP Protection Profile
SAR Security Assurance Requirement
SFP Security Function Policy
SFR Security Functional Requirement
SSH Secure Shell
ST Security Target
TOE Target of Evaluation
TLS Transport Layer Security
TSF TOE security function
TSFI TSF Interface
TSP TOE security policy