Agent presence for self-healing
In one or more examples, there is disclosed a system and method of detecting agent presence for self-healing. An out-of-band monitoring process, such as Intel® AMT, or any process in firmware executing on a co-processor, may monitor one or more processes to determine if one goes down or otherwise meets a security criterion. Crashed processes may be reported to an enterprise security controller (ESC). The ESC may notice trends among affected machines and instruct the machines to take appropriate remedial action, such as booting from a remedial image.
1. An enterprise security controller to provide security to an enterprise, comprising:
a processor and memory;
a network interface; and
executable instructions encoded within the memory to provide a security engine configured to
receive via the network interface a plurality of reports of security events from a plurality of client devices, the security events indicating that a plurality of malfunctions or crashes of an application has occurred across the plurality of client devices of the enterprise, wherein the application is indicated in a defined security policy;
perform a determination from the reports that the plurality of malfunctions or crashes occurred within a predetermined time; and
instruct, based on the determination, at least one of the plurality of client devices to take a remedial action, wherein the remedial action comprises causing the at least one of the plurality of client devices to boot the at least one of the plurality of client devices from a remediation image.
2. The enterprise security controller of claim 1 , wherein one of the security events comprises detecting that a monitored process uses excessive system resources.
3. The enterprise security controller of claim 1 , wherein one of the security events comprises detecting a change of process identifier for a monitored process.
4. The enterprise security controller of claim 1 , wherein one of the security events comprises detecting removal or uninstallation of a monitored application.
5. The enterprise security controller of claim 1 , wherein one of the security events comprises detecting a system resource leak from a monitored process.
6. The enterprise security controller of claim 1 , wherein the remedial action comprises causing the at least one of the plurality of client devices to reinstall a monitored application.
7. The enterprise security controller of claim 1 , wherein the remedial action comprises causing the at least one of the plurality of client devices to restart a monitored process.
8. The enterprise security controller of claim 1 , wherein the remedial action comprises causing the at least one of the plurality of client devices to run a diagnostic command.
9. The enterprise security controller of claim 1 , wherein the remedial action comprises causing the at least one of the plurality of client devices to reimage the at least one of the plurality of client devices with a clean operating system image.
10. One or more non-transitory, computer-readable mediums having stored thereon executable instructions to instruct a processor to provide a security server engine to provide security to an enterprise, the security server engine configured to:
receive via a network interface a plurality of reports of security events from a plurality of client devices, the security events indicating that a plurality of malfunctions or crashes of an application has occurred across the plurality of client devices of the enterprise, wherein the application is indicated in a defined security policy;
perform a determination from the reports that the plurality of malfunctions or crashes has occurred within a designated period; and
instruct, based on the determination, at least one of the plurality of client devices to take a remedial action, wherein the remedial action comprises causing the at least one of the plurality of client devices to boot the at least one of the plurality of client devices from a remediation image.
11. The one or more computer-readable mediums of claim 10 , wherein one of the security events comprises detecting that a monitored process uses excessive system resources.
12. The one or more computer-readable mediums of claim 10 , wherein one of the security events comprises detecting a change of process identifier for a monitored process.
13. The one or more computer-readable mediums of claim 10 , wherein one of the security events comprises detecting removal or uninstallation of a monitored application.
14. The one or more computer-readable mediums of claim 10 , wherein one of the security events comprises detecting an error in a monitored process.
15. The one or more computer-readable mediums of claim 10 , wherein one of the security events comprises detecting a system resource leak from a monitored process.
16. The one or more computer-readable mediums of claim 10 , further comprising:
a machine learning engine configured to detect and avoid false positives based on historical data.
17. The enterprise security controller of claim 1 , wherein the security engine is further configured to enforce a security policy.
18. The enterprise security controller of claim 1 , further comprising:
a machine learning engine configured to receive feedback to determine whether the at least one of the plurality of client devices hosted malware.
19. The enterprise security controller of claim 1 , wherein the remedial action comprises placing at least one of the plurality of client devices in a protected subnetwork.
20. A method, comprising:
receiving via a network interface a plurality of reports of security events from a plurality of client devices of an enterprise, the security events indicating that a plurality of malfunctions or crashes of an application has occurred across the plurality of client devices, wherein the application is indicated in a defined security policy;
performing, via a security engine within an enterprise security controller (ESC), a determination from the reports that the plurality of malfunctions or crashes occurred within a predetermined time; and
instructing, based on the determination, via the security engine, at least one of the plurality of client devices to take a remedial action, wherein the remedial action comprises causing the at least one of the plurality of client devices to boot the at least one of the plurality of client devices from a remediation image.
21. The method of claim 20 , wherein one of the security events comprises detecting that a monitored process uses excessive system resources, detecting a change of process identifier for a monitored process, detecting removal or uninstallation of a monitored application, detecting an error in a monitored process, or detecting a system resource leak from a monitored process.
22. The method of claim 20 , wherein the remedial action comprises causing the at least one of the plurality of client devices to reinstall a monitored application, causing the at least one of the plurality of client devices to restart a monitored process, causing the at least one of the plurality of client devices to run a diagnostic command, causing the at least one of the plurality of client devices to reimage the at least one of the plurality of client devices with a clean operating system image, or placing at least one of the plurality of client devices in a protected subnetwork.
23. The method of claim 20 , further comprising:
determining, with a machine learning engine, whether the at least one of the plurality of client devices hosted malware, based on received feedback.