An Information Model for the Monitoring of Network Security Functions (NSF)

According to , the interface provided by a NSF (e.g., FW, IPS, Anti-DDOS, or Anti-Virus) to administrative entities (e.g., NMS, security controller) for configuring security function in the NSF and monitoring the NSF is referred to as a 'I2NSF customer-facing interface'. The monitoring part of it is meant to monitor the NSF e.g. events, alerts, alarms, logs, counters. The monitoring of the NSF plays a very important role in the overall security framework if done in a timely and comprehensive way. The monitoring information generated by a NSF could very well be an early indication of malicious activity, or anomalous behavior, or a potential sign of denial of service attacks. This draft proposes a comprehensive NSF monitoring information model that provide visibility into NSFs. This document will not go into the design details of NSF-facing interface. Instead, this draft is focused on specifying the information that a NSF needs to provide in the monitoring part of the NSF-facing interface, as well as its information model. Besides, [I-D.draft-xibassnez-i2nsf-capability ] specifies the information model for the security policy provisioning part of the NSF-facing interface.

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 [RFC2119].

This document uses the terms defined in [I-D.draft-ietf-i2nsf-terminology].

As mentioned earlier, monitoring plays a very critical role in the overall security framework. The monitoring of the NSF provides very valuable information to the security controller in maintaining the provisioned security posture. Besides this, there are various other reasons to monitor the NSFs as listed below: The security administrator could configure a policy that is triggered on a specific event happened in the NSF or the network. The security controller would monitor for the specified event and once it happens, it configures additional security functions as per the policy. The events triggered by NSFs as a result of security policy violation could be used by SIEM to detect any suspicious activity. The events and activity logs from NSFs could be used to build advanced analytics such as behavior and predictive to improve the security posture. The security controller could use events from the NSF for achieving high availability. It could take corrective actions such as restarting a failed NSF, horizontally scaling the NSF etc. The events and activity logs from the NSF could aid in debugging and root cause analysis of an operational issue. The activity logs from the NSF could be used to build historical data for operational and business reasons.

In order to maintain a strong security posture, it is not only necessary to configure NSF security policies but also to continuously monitor NSF by consuming acquirable monitoring data. This enables security admins to assess what is happening in the network timely. It is not possible to block all the internal and external threats based on static security posture but requires a very dynamic posture with constant visibility. This draft defines a set of information elements (and their scope) that can be acquired from NSF and can be used as monitoring data. In essence, these types of monitoring data can be leveraged to support constant visibility on multiple levels of granularity and can be consumed by corresponding functions. The types of monitoring data as ordered below increase in expressiveness by incorporating more information to the semantics of the monitoring data. There are two categories of monitoring data. Information that is produced and emitted by an NSF automatically (published data) and information that is produced and retained by the NSF and has to be collected in intervals (retained data): Published Data: Events: the most generic type of monitoring data that can be emitted by an NSF. An event is defined in [RFC3877] as “something that happens which may be of interest. A fault, a change in status, crossing a threshold, or an external input to the system, for example. In the context of the I2NSF IMM, an event is a representation of a change of state, configuration, or composition of an NSF or an entity (e.g. an endpoint) or an activity (e.g. a PDU flow) that can be observed by the NSF and can be interpreted as a change of state or behavior by the NSF. An event can be created without the use of an I2NSF Condition (declarative guidance) available to the NSF. In the context of I2NSF, in some cases an event can trigger low level I2NSF actions (which constitutes an implicit escalation to alert via primate assessment). Alert: an event that is annotated with a criticality assessment due to non-compliance with I2NSF conditions (declarative guidance) available to an I2NSF Consumer via I2NSF Actions. The Intrusion Detection Message Exchange Format [RFC4765] defines a representation that “systems can use to report alerts about events that they deem suspicious” and also associates a severity (“an estimate of the relative severity of the event”) with the corresponding alert. In the context of the I2NSF IMM, an alert is derived from events that express changes indicating not to conform with declarative guidance (e.g. an exceeded threshold of a value or a pattern or signature found in a PDU stream – typically an I2NSF condition) or due to imperative guidance (e.g. correlation rules applied to streams of multiple events over time or a black-list content of an event matches – typically an I2NSF Policy Rule). An alert is created by an I2NSF Producer with respect to I2NSF conditions (declarative guidance) or imperative guidance available to the I2NSF Producer. An alert does not indicate an immediate impact on operations and are not time-sensitive (but can be escalated to an alarm nevertheless due to persistence via an I2NSF Policy Rule). Alarms: an alarm that is annotated with the assertion of: 1.) having immediate impact on operations, or 2.) a persistence that in non-compliant in respect to I2NSF conditions (declarative guidance), or 3.) a correlation result produced by a I2NSF Service in respect to the result of I2NSF Policy Rules that process alerts. Alarms are time-sensitive and must be reported to the security admin as soon as possible. By processing alarms, the administrator can rapidly locate the root-cause of faults and rapidly deal with the faults to ensure normal operation of the NSFs and avoid NSFs going into unknown state or potentially exposing security vulnerabilities. An analyst can manage the NSF with via I2NSF Policy Rules. The intend of alarms is to highlight only critical information and to avoid continuous combing through large amount alerts (or even events) by analysts. Retained Data: Logs: Logs are information generated by NSF about its operational and informational data, or various events such as user activities, network/traffic status and network activity, etc. Logs are important for debugging, auditing and security forensic. Unlike events, logs do not require an immediate attention from an analyst but may be useful for visibility and retroactive cyber forensic. Hence, logs usually include less structures but potentially more detailed information in regard to events. For example, the NSF could generate a log when due to an I2NSF Policy Rule. Logs can be continuously processed by I2NSF Agent that act as I2NSF Producer and emit events via function specifically tailored to a type of log. Counters: A specific representation of identical state changes that potentially occur in high frequency. Examples include network interface counters (packets, bytes), drop, error counters etc. Counters are useful in debugging and visibility into operational behavior of the NSF. An I2NSF Agent that observes the progress of counters can act as an I2NSF Producer and emit Events in respect to I2NSF Policy Rules.

As per the use cases of NSF monitoring data, the data need to be sent to various consumers based on the needs and requirements. There are multiple things to be considered for exporting this data to needed parties as listed below: Pull-Push model: A set of data could be pushed by a NSF to the needed party or pulled by the needed party from a NSF. A specific data might need both the models at the same time if there are multiple consumers with varying requirements. It really depends upon the need and its usages to the consumer. In general, any alarm is considered to be of great importance and must be sent immediately for any meaningful action so it should be sent using push model but logs are not as critical so could be pulled by the consumer. The I2NSF does not mandate a specific scheme for each data set, it is up to each implementation. Export frequency: The monitoring data could be sent immediately upon generation by a NSF to interested parties or pushed periodically. The frequency of exporting the data depends upon its size and timely usefulness. It is out of the scope of I2NSF and left to each NSF implementation. Authentication: There may be a need for authentication between monitoring data producer (NSF) and consumer to ensure that critical information does not fall into wrong hands. This may be necessary if the NSF directly export data to the consumer outside its admin boundary. The I2NSF does not mandate when and how specific authentication must be done. Subscription method: In order for the consumer to pull the data from NSF or for NSF to push the data to a consumer, there must be a mechanism for consumer to subscribe to the NSF data it is interested in. There are few open source method available and it is up to each implementation to decide the right one. Data transfer mode: The data could be pushed by NSF using a connection-less model that does require a persistent connection or streamed over a persistent connection. It depends upon the requirement of the consumer and the nature of data. A particular set of data can use either or both the mode based on implementation. Transport method: There are lot of transport mechanism such as IP, UDP, TCP. There are also open source implementations for specific set of data such as systems counter. The I2NSF does not mandate any specific method for a given data set, it is up to each implementation.

As explained in the above section, there is a wealth of data available from the NSF that can be monitored. Firstly, there must be some general information with each monitoring message sent from an NSF that helps consumer in identifying meta data with that message, which are listed as below: message_version: Indicate the version of the data format and is a two-digit decimal numeral starting from 01 message_type: Event, Alert, Alarm, Log, Counter, etc time_stamp: Indicate the time when the message is generated vendor_name: The name of the NSF vendor NSF_name: The name (or IP) of the NSF generating the message Module_name: The module name outputting the message Severity: Indicates the level of the logs. There are total eight levels, from 0 to 7. The smaller the numeral is, the higher the severity is.

This section covers the additional information associated with the system messages. The extended information model is only for the structured data such as alarm. Any unstructured data is specified with basic information model only. [Editors' note]: This section remains the same as -02 version, although the classification of the monitoring data has been changed from -02 version. The new inconsistency will be addressed in next verion.

The following information should be included in a Memory Alarm: event_name: 'MEM_USAGE_ALARM' module_name: Indicate the NSF module responsible for generating this alarm usage: specifies the amount of memory used threshold: The threshold triggering the alarm severity: The severity of the alarm such as critical, high, medium, low message: 'The memory usage exceeded the threshold'

The following information should be included in a CPU Alarm: event_name: 'CPU_USAGE_ALARM' usage: Specifies the amount of CPU used threshold: The threshold triggering the event severity: The severity of the alarm such as critical, high, medium, low message: 'The CPU usage exceeded the threshold'

The following information should be included in a Disk Alarm: event_name: 'DISK_USAGE_ALARM' usage: Specifies the amount of disk space used threshold: The threshold triggering the event severity: The severity of the alarm such as critical, high, medium, low message: 'The disk usage exceeded the threshold'

The following information should be included in a Hardware Alarm: event_name: 'HW_FAILURE_ALARM' component_name: Indicate the HW component responsible for generating this alarm threshold: The threshold triggering the alarm severity: The severity of the alarm such as critical, high, medium, low message: 'The HW component has failed or degraded'

The following information should be included in a Interface Alarm: event_name: 'IFNET_STATE_ALARM' interface_Name: The name of interface interface_state: 'UP', 'DOWN', 'CONGESTED' threshold: The threshold triggering the event severity: The severity of the alarm such as critical, high, medium, low message: 'Current interface state'

The following information should be included in this event: event_name: 'ACCESS_DENIED' user: Name of a user group: Group to which a user belongs login_ip_address: Login IP address of a user authentication_mode: User authentication mode. e.g., Local Authentication, Third-Party Server Authentication, Authentication Exemption, SSO Authentication message: 'access denied'

The following information should be included in this event: event_name: 'CONFIG_CHANGE' user: Name of a user group: Group to which a user belongs login_ip_address: Login IP address of a user authentication_mode: User authentication mode. e.g., Local Authentication, Third-Party Server Authentication, Authentication Exemption, SSO Authentication message: 'Configuration modified'

Access logs record administrators' login, logout, and operations on the device. By analyzing them, security vulnerabilities can be identified. The following information should be included in operation report: Administrator: Administrator that operates on the device login_ip_address: IP address used by an administrator to log in login_mode: Specifies the administrator logs in mode e.g. root, user operation_type: The operation type that the administrator execute, e.g., login, logout, configuration, etc result: Command execution result content: Operation performed by an administrator after login.

Running reports record the device system's running status, which is useful for device monitoring. The following information should be included in running report: system_status: The current system's running status CPU_usage: Specifies the CPU usage memory_usage: Specifies the memory usage disk_usage: Specifies the disk usage disk_left: Specifies the available disk space left session_number: Specifies total concurrent sessions process_number: Specifies total number of system processes in_traffic_rate: The total inbound traffic rate in pps out_traffic_rate: The total outbound traffic rate in pps in_traffic_speed: The total inbound traffic speed in bps out_traffic_speed: The total outbound traffic speed in bps

User activity logs provide visibility into users' online records (such as login time, online/lockout duration, and login IP addresses) and the actions users perform. User activity reports are helpful to identify exceptions during user login and network access activities. user: Name of a user group: Group to which a user belongs login_ip_address: Login IP address of a user authentication_mode: User authentication mode. e.g., Local Authentication, Third-Party Server Authentication, Authentication Exemption, SSO Authentication access_mode: User access mode. e.g., PPP, SVN, LOCAL online_duration: Online duration lockout_duration: Lockout duration type: User activities. e.g., Successful User Login, Failed Login attempts, User Logout, Successful User Password Change, Failed User Password Change, User Lockout, User Unlocking, Unknown cause: Cause of a failed user activity

Interface counters provide visibility into traffic into and out of NSF, bandwidth usage. interface_name: Network interface name configured in NSF in_total_traffic_pkts: Total inbound packets out_total_traffic_pkts: Total outbound packets in_total_traffic_bytes: Total inbound bytes out_total_traffic_bytes: Total outbound bytes in_drop_traffic_pkts: Total inbound drop packets out_drop_traffic_pkts: Total outbound drop packets in_drop_traffic_bytes: Total inbound drop bytes out_drop_traffic_bytes: Total outbound drop bytes in_traffic_ave_rate: Inbound traffic average rate in pps in_traffic_peak_rate: Inbound traffic peak rate in pps in_traffic_ave_speed: Inbound traffic average speed in bps in_traffic_peak_speed: Inbound traffic peak speed in bps out_traffic_ave_rate: Outbound traffic average rate in pps out_traffic_peak_rate: Outbound traffic peak rate in pps out_traffic_ave_speed: Outbound traffic average speed in bps out_traffic_peak_speed: Outbound traffic peak speed in bps.

The following information should be included in a DDoS Event: event_name: 'SEC_EVENT_DDoS' sub_attack_type: Any one of Syn flood, ACK flood, SYN-ACK flood, FIN/RST flood, TCP Connection flood, UDP flood, Icmp flood, HTTPS flood, HTTP flood, DNS query flood, DNS reply flood, SIP flood, and etc. dst_ip: The IP address of a victum under attack dst_port: The port numbers that the attrack traffic aims at. start_time: The time stamp indicating when the attack started end_time: The time stamp indicating when the attack ended. If the attack is still undergoing when sending out the alarm, this field can be empty. attack_rate: The PPS of attack traffic attack_speed: the bps of attack traffic rule_id: The ID of the rule being triggered rule_name: The name of the rule being triggered profile: Security profile that traffic matches.

The following information should be included in a Session Table Event: event_name: 'SESSION_USAGE_HIGH' current: The number of concurrent sessions max: The maximum number of sessions that the session table can support threshold: The threshold triggering the event message: 'The number of session table exceeded the threshold'

The following information should be included in a Virus Event: event_Name: 'SEC_EVENT_VIRUS' virus_type: Type of the virus, e.g., trojan, worm, macro Virus type virus_name dst_ip: The destination IP address of the packet where the virus is found src_ip: The source IP address of the packet where the virus is found src_port: The source port of the packet where the virus is found dst_port: The destination port of the packet where the virus is found src_zone: The source security zone of the packet where the virus is found dst_zone: The destination security zone of the packet where the virus is found file_type: The type of the file where the virus is hided within file_name: The name of the file where the virus is hided within virus_info: The brief introduction of virus raw_info: The information describing the packet triggering the event. rule_id: The ID of the rule being triggered rule_name: The name of the rule being triggered profile: Security profile that traffic matches.

The following information should be included in a Intrustion Event: event_name: The name of event: 'SEC_EVENT_Intrusion' sub_attack_type: Attack type, e.g., brutal force, buffer overflow src_ip: The source IP address of the packet dst_ip: The destination IP address of the packet src_port:The source port number of the packet dst_port: The destination port number of the packet src_zone: The source security zone of the packet dst_zone: The destination security zone of the packet protocol: The employed transport layer protocol, e.g.,TCP, UDP app: The employed application layer protocol, e.g.,HTTP, FTP rule_id: The ID of the rule being triggered rule_name: The name of the rule being triggered profile: Security profile that traffic matches intrusion_info: Simple description of intrusion raw_info: The information describing the packet triggering the event.

The following information should be included in a Botnet Event: event_name: the name of event: 'SEC_EVENT_Botnet' botnet_name: The name of the detected botnet src_ip: The source IP address of the packet dst_ip: The destination IP address of the packet src_port: The source port number of the packet dst_port: The destination port number of the packet src_zone: The source security zone of the packet dst_zone: The destination security zone of the packet protocol: The employed transport layer protocol, e.g.,TCP, UDP app: The employed application layer protocol, e.g.,HTTP, FTP role: The role of the communicating parties within the botnet: the packet from zombie host to the attacker The packet from the attacker to the zombie host The packet from the IRC/WEB server to the zombie host The packet from the zombie host to the IRC/WEB server The packet from the attacker to the IRC/WEB server The packet from the IRC/WEB server to the attacker The packet from the zombie host to the victim botnet_info: Simple description of Botnet rule_id: The ID of the rule being triggered rule_name: The name of the rule being triggered profile: Security profile that traffic matches raw_info: The information describing the packet triggering the event.

The following information should be included in a Web Attack Alarm: event_name: the name of event: 'SEC_EVENT_WebAttack' sub_attack_type: Concret web attack type, e.g., sql injection, command injection, XSS, CSRF src_ip: The source IP address of the packet dst_ip: The destination IP address of the packet src_port: The source port number of the packet dst_port: The destination port number of the packet src_zone: The source security zone of the packet dst_zone: The destination security zone of the packet req_method: The method of requirement. For instance, 'PUT' or 'GET' in HTTP req_url: Requested URL url_category: Matched URL category filtering_type: URL filtering type, e.g., Blacklist, Whitelist, User-Defined, Predefined, Malicious Category, Unknown rule_id: The ID of the rule being triggered rule_name: The name of the rule being triggered profile: Security profile that traffic matches.

Besides the fields in an DDoS Alarm, the following information should be included in a DDoS Logs: attack_type: DDoS attack_ave_rate: The average pps of the attack traffic within the recorded time attack_ave_speed: The average bps of the attack traffic within the recorded time attack_pkt_num: The number attack packets within the recorded time attack_src_ip: The source IP addresses of attack traffics. If there are a large amount of IP addresses, then pick a certain number of resources according to different rules. action: Actions against DDoS attacks, e.g., Allow, Alert, Block, Discard, Declare, Block-ip, Block-service.

Besides the fields in an Virus Alarm, the following information should be included in a Virus Logs: attack_type: Virus protocol: The transport layer protocol app: The name of the application layer protocol times: The time of detecting the virus action: The actions dealing with the virus, e.g., alert, block os: The OS that the virus will affect, e.g., all, android, ios, unix, windows

Besides the fields in an Intrusion Alarm, the following information should be included in a Intrusion Logs: attack_type: Intrusion times: The times of intrusions happened in the recorded time os: The OS that the intrusion will affect, e.g., all, android, ios, unix, windows action: The actions dealing with the intrusions, e.g., e.g., Allow, Alert, Block, Discard, Declare, Block-ip, Block-service attack_rate: NUM the pps of attack traffic attack_speed: NUM the bps of attack traffic

Besides the fields in an Botnet Alarm, the following information should be included in a Botnet Logs: attack_type: Botnet botnet_pkt_num:The number of the packets sent to or from the detected botnet action: The actions dealing with the detected packets, e.g., Allow, Alert, Block, Discard, Declare, Block-ip, Block-service, etc os: The OS that the attack aiming at, e.g., all, android, ios, unix, windows, etc.

DPI Logs provide statistics on uploaded and downloaded files and data, sent and received emails, and alert and block records on websites. It's helpful to learn risky user behaviors and why access to some URLs is blocked or allowed with an alert record. type: DPI action types. e.g., File Blocking, Data Filtering, Application Behavior Control file_name: The file name file_type: The file type src_zone: Source security zone of traffic dst_zone: Destination security zone of traffic src_region: Source region of the traffic dst_region: Destination region of the traffic src_ip: Source IP address of traffic src_user: User who generates traffic dst_ip: Destination IP address of traffic src_port: Source port of traffic dst_port: Destination port of traffic protocol: Protocol type of traffic app: Application type of traffic policy_id: Security policy id that traffic matches policy_name: Security policy name that traffic matches action: Action defined in the file blocking rule, data filtering rule, or application behavior control rule that traffic matches.

Vulnerability scanning logs record the victim host and its related vulnerability information that should to be fixed. the following information should be included in the report: victim_ip: IP address of the victim host which has vulnerabilities vulnerability_id: The vulnerability id vulnerability_level: The vulnerability level. e.g., high, middle, low OS: The operating system of the victim host service: The service which has vulnerabillity in the victim host protocol: The protocol type. e.g., TCP, UDP port: The port number vulnerability_info: The information about the vulnerability fix_suggestion: The fix suggestion to the vulnerability.

Besides the fields in an Web Attack Alarm, the following information should be included in a Web Attack Report: attack_type: Web Attack rsp_code: Response code req_clientapp: The client application req_cookies: Cookies req_host: The domain name of the requested host raw_info: The information describing the packet triggering the event.

Firewall counters provide visibility into traffic signatures, bandwidth usage, and how the configured security and bandwidth policies have been applied. src_zone: Source security zone of traffic dst_zone: Destination security zone of traffic src_region: Source region of the traffic dst_region: Destination region of the traffic src_ip: Source IP address of traffic src_user: User who generates traffic dst_ip: Destination IP address of traffic src_port: Source port of traffic dst_port: Destination port of traffic protocol: Protocol type of traffic app: Application type of traffic policy_id: Security policy id that traffic matches policy_name: Security policy name that traffic matches in_interface: Inbound interface of traffic out_interface: Outbound interface of traffic total_traffic: Total traffic volume in_traffic_ave_rate: Inbound traffic average rate in pps in_traffic_peak_rate: Inbound traffic peak rate in pps in_traffic_ave_speed: Inbound traffic average speed in bps in_traffic_peak_speed: Inbound traffic peak speed in bps out_traffic_ave_rate: Outbound traffic average rate in pps out_traffic_peak_rate: Outbound traffic peak rate in pps out_traffic_ave_speed: Outbound traffic average speed in bps out_traffic_peak_speed: Outbound traffic peak speed in bps.

Policy Hit Counters record the security policy that traffic matches and its hit count. It can check if policy configurations are correct. src_zone: Source security zone of traffic dst_zone: Destination security zone of traffic src_region: Source region of the traffic dst_region: Destination region of the traffic src_ip: Source IP address of traffic src_user: User who generates traffic dst_ip: Destination IP address of traffic src_port: Source port of traffic dst_port: Destination port of traffic protocol: Protocol type of traffic app: Application type of traffic policy_id: Security policy id that traffic matches policy_name: Security policy name that traffic matches hit_times: The hit times that the security policy matches the specified traffic.

This document makes no request of IANA. Note to RFC Editor: this section may be removed on publication as an RFC.

The monitoring information of NSF should be protected by the secure communication channel, to ensure its confidentiality and integrity. In another side, the NSF and security controller can all be faked, which lead to undesireable results, i.e., leakage of NSF's important operational information, faked NSF sending false information to mislead security controller. The mutual authentication is essential to protected against this kind of attack. The current mainstream security technologies (i.e., TLS, DTLS, IPSEC, X.509 PKI) can be employed approriately to provide the above security functions. In addition, to defend against the DDoS attack caused by a lot of NSFs sending massive monitoring information to the security controller, the rate limiting or similar mechanisms should be considered in NSF and security controller, whether in advance or just in the process of DDoS attack.