Network Device Management Security Requirements Guide

Successful identification and authentication must not automatically give an entity full access to a network device or security domain. The lack of authorization-based access control could result in the immediate compromise and unauthorized access to sensitive information. All DoD systems must be properly configured to incorporate access control methods that do not rely solely on authentication for authorized access. Authorization is the process of determining whether an entity, once authenticated, is permitted to access a specific asset or set of resources. Information systems use access control policies and enforcement mechanisms to implement this requirement. Authorization procedures and controls must be implemented to ensure each authenticated entity also has a validated and current authorization. Some network devices are pre-configured with security groups. Other network devices enable operators to create custom security groups with custom permissions. For example, an ISSM may require read-only access to audit the network device. Operators may create an audit security group, define permissions and access levels for members of the group, and then assign the ISSM’s user persona to the audit security group. This is still considered privileged access, but the ISSM’s security group is more restrictive than the network administrator’s security group. Network devices that rely on AAA brokers for authentication and authorization services may need to identify the available security groups or access levels available on the network devices and convey that information to the AAA operator. Once the AAA broker identifies the user persona on the centralized directory service, the user’s security group memberships can be retrieved. The AAA operator may need to create a mapping that links target security groups from the directory service to the appropriate security groups or access levels on the network device. Once these mappings are configured, authorizations can happen dynamically, based on each user’s directory service group membership.

In order to prevent unauthorized connection of devices, unauthorized transfer of information, or unauthorized tunneling (i.e., embedding of data types within data types), organizations must disable unused or unnecessary physical and logical ports/protocols on information systems. Network devices are capable of providing a wide variety of functions and services. Some of the functions and services provided by default may not be necessary to support essential organizational operations. Additionally, it is sometimes convenient to provide multiple services from a single component (e.g., email and web services); however, doing so increases risk over limiting the services provided by any one component. To support the requirements and principles of least functionality, the network device must support the organizational requirements providing only essential capabilities and limiting the use of ports, protocols, and/or services to only those required, authorized, and approved. Some network devices have capabilities enabled by default; if these capabilities are not necessary, they must be disabled. If a particular capability is used, then it must be documented and approved.

Passwords need to be protected at all times, and encryption is the standard method for protecting passwords. If passwords are not encrypted, they can be plainly read (i.e., clear text) and easily compromised. Network devices must enforce cryptographic representations of passwords when storing passwords in databases, configuration files, and log files. Passwords must be protected at all times; using a strong one-way hashing encryption algorithm with a salt is the standard method for providing a means to validate a password without having to store the actual password. Performance and time required to access are factors that must be considered, and the one way hash is the most feasible means of securing the password and providing an acceptable measure of password security. If passwords are stored in clear text, they can be plainly read and easily compromised. In many instances, verifying the user knows a password is performed using a password verifier. In its simplest form, a password verifier is a computational function that is capable of creating a hash of a password and determining if the value provided by the user matches the stored hash.

Passwords need to be protected at all times, and encryption is the standard method for protecting passwords. If passwords are not encrypted, they can be plainly read (i.e., clear text) and easily compromised. Network devices can accomplish this by making direct function calls to encryption modules or by leveraging operating system encryption capabilities.

To prevent the compromise of authentication information such as passwords during the authentication process, the feedback from the network device must not provide any information that would allow an unauthorized user to compromise the authentication mechanism. Obfuscation of user-provided information when typed into the system is a method used in addressing this risk. For example, displaying asterisks when a user types in a password is an example of obscuring feedback of authentication information.

Unapproved mechanisms that are used for authentication to the cryptographic module are not validated and therefore cannot be relied upon to provide confidentiality or integrity, and DoD data may be compromised. Network devices utilizing encryption are required to use FIPS-compliant mechanisms for authenticating to cryptographic modules. FIPS 140-2 is the current standard for validating that mechanisms used to access cryptographic modules utilize authentication that meets DoD requirements. However, authentication algorithms must configure security processes to use only FIPS-approved and NIST-recommended authentication algorithms.

Terminating an idle session within a short time period reduces the window of opportunity for unauthorized personnel to take control of a management session enabled on the console or console port that has been left unattended. In addition, quickly terminating an idle session will also free up resources committed by the managed network element. Terminating network connections associated with communications sessions includes, for example, de-allocating associated TCP/IP address/port pairs at the operating system level, or de-allocating networking assignments at the application level if multiple application sessions are using a single, operating system-level network connection. This does not mean that the device terminates all sessions or network access; it only ends the inactive session and releases the resources associated with that session.

This requirement is intended to address the confidentiality and integrity of system information at rest (e.g., network device rule sets) when it is located on a storage device within the network device or as a component of the network device. This protection is required to prevent unauthorized alteration, corruption, or disclosure of information when not stored directly on the network device. Files on the network device or on removable media used by the device must have their permissions set to allow read or write access to those accounts that are specifically authorized to access or change them. Note that different administrative accounts or roles will have varying levels of access. File permissions must be set so that only authorized administrators can read or change their contents. Whenever files are written to removable media and the media removed from the device, the media must be handled appropriately for the classification and sensitivity of the data stored on the device.

Preventing non-privileged users from executing privileged functions mitigates the risk that unauthorized individuals or processes may gain unnecessary access to information or privileges. Privileged functions include, for example, establishing accounts, performing system integrity checks, or administering cryptographic key management activities. Non-privileged users are individuals that do not possess appropriate authorizations.

Unapproved mechanisms that are used for authentication to the cryptographic module are not verified and therefore cannot be relied upon to provide confidentiality or integrity, and DoD data may be compromised. Nonlocal maintenance and diagnostic activities are those activities conducted by individuals communicating through a network, either an external network (e.g., the Internet) or an internal network. Currently, HMAC is the only FIPS-approved algorithm for generating and verifying message/data authentication codes in accordance with FIPS 198-1. Products that are FIPS 140-2 validated will have an HMAC that meets specification; however, the option must be configured for use as the only message authentication code used for authentication to cryptographic modules. Separate requirements for configuring applications and protocols used by each application (e.g., SNMPv3, SSHv2, NTP, HTTPS, and other protocols and applications that require server/client authentication) are required to implement this requirement. Where SSH is used, the SSHv2 protocol suite is required because it includes Layer 7 protocols such as SCP and SFTP, which can be used for secure file transfers.

This requires the use of secure protocols instead of their unsecured counterparts, such as SSH instead of telnet, SCP instead of FTP, and HTTPS instead of HTTP. If unsecured protocols (lacking cryptographic mechanisms) are used for sessions, the contents of those sessions will be susceptible to eavesdropping, potentially putting sensitive data (including administrator passwords) at risk of compromise and potentially allowing hijacking of maintenance sessions.

Centralized management of authentication settings increases the security of remote and nonlocal access methods. This control is particularly important protection against the insider threat. With robust centralized management, audit records for administrator account access to the organization's network devices can be more readily analyzed for trends and anomalies. The alternative method of defining administrator accounts on each device exposes the device configuration to remote access authentication attacks and system administrators with multiple authenticators for each network device.

The aggregation of log data kept on a syslog server can be used to detect attacks and trigger an alert to the appropriate security personnel. The stored log data can used to detect weaknesses in security that enable the network IA team to find and address these weaknesses before breaches can occur. Reviewing these logs, whether before or after a security breach, are important in showing whether someone is an internal employee or an outside threat.

Network devices running an unsupported operating system lack current security fixes required to mitigate the risks associated with recent vulnerabilities.

Multi-factor authentication (MFA) is when two or more factors are used to confirm the identity of an individual who is requesting access to digital information resources. Valid factors include something the individual knows (e.g., username and password), something the individual has (e.g., a smartcard or token), or something the individual is (e.g., a fingerprint or biometric). Legacy information system environments only use a single factor for authentication, typically a username and password combination. Although two pieces of data are used in a username and password combination, this is still considered single factor because an attacker can obtain access simply by learning what the user knows. Common attacks against single-factor authentication are attacks on user passwords. These attacks include brute force password guessing, password spraying, and password credential stuffing. MFA, along with strong user account hygiene, helps mitigate against the threat of having account passwords discovered by an attacker. Even in the event of a password compromise, with MFA implemented and required for interactive login, the attacker still needs to acquire something the user has or replicate a piece of user’s biometric digital presence. Private industry recognizes and uses a wide variety of MFA solutions. However, DoD public key infrastructure (PKI) is the only prescribed method approved for DoD organizations to implement MFA. For authentication purposes, centralized DoD certificate authorities (CA) issue PKI certificate key pairs (public and private) to individuals using the prescribed x.509 format. The private certificates that have been generated by the issuing CA are downloaded and saved to smartcards which, within DoD, are referred to as common access cards (CAC) or personal identity verification (PIV) cards. This happens at designated DoD badge facilities. The CA maintains a record of the corresponding public keys for use with PKI-enabled environments. Privileged user smartcards, or “alternate tokens”, function in the same manner, so this requirement applies to all interactive user sessions (authorized and privileged users). Note: This requirement is used in conjunction with the use of a centralized authentication server (e.g., AAA, RADIUS, LDAP), a separate but equally important requirement. The MFA configuration of this requirement provides identification and the first phase of authentication (the challenge and validated response, thereby confirming the PKI certificate that was presented by the user). The centralized authentication server will provide the second phase of authentication (the digital presence of the PKI ID as a valid user in the requested security domain) and authorization. The centralized authentication server will map validated PKI identities to valid user accounts and determine access levels for authenticated users based on security group membership and role. In cases where the centralized authentication server is not utilized by the network device for user authorization, the network device must map the authenticated identity to the user account for PKI-based authentication.

Once issued by a DoD certificate authority (CA), public key infrastructure (PKI) certificates are typically valid for 3 years or shorter within the DoD. However, there are many reasons a certificate may become invalid before the prescribed expiration date. For example, an employee may leave or be terminated and still possess the smartcard on which the PKI certificates were stored. Another example is that a smartcard containing PKI certificates may become lost or stolen. A more serious issue could be that the CA or server which issued the PKI certificates has become compromised, thereby jeopardizing every certificate keypair that was issued by the CA. These examples of revocation use cases and many more can be researched further using Internet cybersecurity resources. PKI user certificates presented as part of the identification and authentication criteria (e.g., DoD PKI as multi-factor authentication [MFA]) must be checked for validity by network devices. For example, valid PKI certificates are digitally signed by a trusted DoD certificate authority (CA). Additionally, valid PKI certificates are not expired, and valid certificates have not been revoked by a DoD CA. Network devices can verify the validity of PKI certificates by checking with an authoritative CA. One method of checking the status of PKI certificates is to query databases referred to as certificate revocation lists (CRL). These are lists which are published, updated, and maintained by authoritative DoD CAs. For example, once certificates are expired or revoked, issuing CAs place the certificates on a certificate revocation list (CRL). Organizations can download these lists periodically (i.e. daily or weekly) and store them locally on the devices themselves or even onto another nearby local enclave resource. Storing them locally ensures revocation status can be checked even if Internet connectivity is severed at the enclave’s point of presence (PoP). However, CRLs can be rather large in storage size and further, the use of CRLs can be rather taxing on some computing resources. Another method of validating certificate status is to use the online certificate status protocol (OCSP). Using OCSP, a requestor (i.e. the network device which the user is trying to authenticate to) sends a request to an authoritative CA challenging the validity of a certificate that has been presented for identification and authentication. The CA receives the request and sends a digitally signed response indicating the status of the user’s certificate as valid, revoked, or unknown. Network devices should only allow access for responses that indicate the certificates presented by the user were considered valid by an approved DoD CA. OCSP is the preferred method because it is fast, provides the most current status, and is lightweight.

Without mapping the PKI certificate to a unique user account, the ability to determine the identities of individuals or the status of their non-repudiation is considerably impacted during forensic analysis. A strength of using PKI as MFA is that it can help ensure only the assigned individual is using their associated user account. This can only be accomplished if the network device is configured to enforce the relationship which binds PKI certificates to unique user accounts. Local accounts (accounts created, stored, and maintained locally on the network device) should be avoided in lieu of using a centrally managed directory service. Local accounts empower the same workgroup who will be operating the network infrastructure to also control and manipulate access methods, thus creating operational autonomy. This undesirable approach breaks the concept of separation of duties. Additionally, local accounts are susceptible to poor cyber hygiene because they create another user database that must be maintained by the operator, whose primary focus is on running the network. Such examples of poor hygiene include dormant accounts that are not disabled or deleted, employees who have left the organization but whose accounts are still present, periodic password and hash rotation, password complexity shortcomings, increased exposure to insider threat, etc. For reasons such as this, local users on network devices are frequently the targets of cyber-attacks. Instead, organizations should explore examples of centrally managed account services. These examples include the implementation of AAA concepts like the use of external RADIUS and LDAP directory service brokers.