Intrusion detection in micro-services through container telemetry and behavior modeling
An intrusion detection system (IDS) for a micro-services environment identifies attacks in substantially real-time and at a container-level. In this approach, behavior models are generated from container images using a binary analysis. A behavior model is a graph data structure having nodes and edges, wherein an edge represents a system call made by at least one process represented as a node in the graph data structure. The model is co-located with a running container, thereby enabling detection of anomalies as the container executes in a container environment on a hardware node. A per-container IDS function is instantiated by checking whether system call telemetry generated by an image's running container satisfies the associated behavior model that has been generated for the container image. If the telemetry indicates activity that deviates from the behavior model, an automated action is then initiated to attempt to address the attack, preferably while it is in progress.
1. A method for intrusion detection in a run-time container environment, comprising:
deploying a behavior model in association with a container executing in the run-time container environment, the behavior model having been generated from a container image and one or more library dependencies of the container image, the behavior model being a graph data structure having a set of nodes, and a set of edges, wherein a node represents one of: a process, a file and a network socket, and wherein an edge represents a system call made by at least one process represented in the graph data structure;
as the container image executes in the container, receiving system call telemetry;
responsive to receipt of the telemetry, determining whether the container image is executing in a manner inconsistent with the behavior model, thereby indicating an anomaly; and
upon a determination that the container image is executing in a manner inconsistent with the behavior model, taking an automated action to attempt to address the anomaly.
2. The method as described in claim 1 further including generating the behavior model for the container image by identifying the library dependencies.
3. The method as described in claim 2 wherein the behavior model is generated by a binary analysis that determines what library functions a given binary calls.
4. The method as described in claim 3 wherein the binary analysis determines what library functions a given binary calls by examining an inter-procedural call graph and finding a set of system calls reachable from the call graph.
5. The method as described in claim 3 wherein the binary analysis generates the behavior model by micro-executing code reachable from an entrypoint of the container image.
6. The method as described in claim 1 wherein the behavior model is instantiated in a hardware node in a Container Orchestration Engine (COE), and wherein the container image executes as a micro-service.
7. The method as described in claim 6 , further including enriching the behavior model based on information representing one or more valid system calls seen by one or more other containers running on the hardware node.
8. The method as described in claim 1 , wherein determining whether the container image is executing in a manner inconsistent with the behavior model is carried out continuously and the automated action is one of: mitigation, notification, sandboxing and logging.
9. Apparatus, comprising:
at least one hardware processor;
computer memory holding computer program instructions executed by the at least one hardware processor to perform intrusion detection in association with a container environment, the computer program instructions comprising program code configured to:
deploy a behavior model in association with a container executing in the run-time container environment, the behavior model having been generated from a container image and one or more library dependencies of the container image, the behavior model being a graph data structure having a set of nodes, and a set of edges, wherein a node represents one of: a process, a file and a network socket, and wherein an edge represents a system call made by at least one process represented in the graph data structure;
as the container image executes in the container, receive system call telemetry;
responsive to receipt of the telemetry, determine whether the container image is executing in a manner inconsistent with the behavior model, thereby indicating an anomaly; and
upon a determination that the container image is executing in a manner inconsistent with the behavior model, take an automated action to attempt to address the anomaly.
10. The apparatus as described in claim 9 wherein the computer program code is further configured to generate the behavior model for the container image by identifying the library dependencies.
11. The apparatus as described in claim 10 wherein the behavior model is generated by a binary analysis that determines what library functions a given binary calls.
12. The apparatus as described in claim 11 wherein the binary analysis determines what library functions a given binary calls by examining an inter-procedural call graph and finding a set of system calls reachable from the call graph.
13. The apparatus as described in claim 11 wherein the binary analysis generates the behavior model by micro-executing code reachable from an entrypoint of the container image.
14. The apparatus as described in claim 9 wherein the behavior model is instantiated in a hardware node in a Container Orchestration Engine (COE), and wherein the container image executes as a micro-service.
15. The apparatus as described in claim 14 , wherein the program code is further configured to enrich the behavior model based on information representing one or more valid system calls seen by one or more other containers running on the hardware node.
16. The apparatus as described in claim 9 , wherein the program code configured to determine whether the container image is executing in a manner inconsistent with the behavior model executes continuously and the automated action is one of: mitigation, notification, sandboxing and logging.
17. A computer program product in a non-transitory computer-readable medium for use in a data processing system, the computer program product holding computer program instructions executed by the data processing system to perform intrusion detection in association with a container environment, the computer program instructions comprising program code configured to:
deploy a behavior model in association with a container executing in the run-time container environment, the behavior model having been generated from a container image and one or more library dependencies of the container image, the behavior model being a graph data structure having a set of nodes, and a set of edges, wherein a node represents one of: a process, a file and a network socket, and wherein an edge represents a system call made by at least one process represented in the graph data structure;
as the container image executes in the container, receive system call telemetry;
responsive to receipt of the telemetry, determine whether the container image is executing in a manner inconsistent with the behavior model, thereby indicating an anomaly; and
upon a determination that the container image is executing in a manner inconsistent with the behavior model, take an automated action to attempt to address the anomaly.
18. The computer program product as described in claim 17 wherein the computer program code is further configured to generate the behavior model for the container image by identifying the library dependencies.
19. The computer program product as described in claim 18 wherein the behavior model is generated by a binary analysis that determines what library functions a given binary calls.
20. The computer program product as described in claim 19 wherein the binary analysis determines what library functions a given binary calls by examining an inter-procedural call graph and finding a set of system calls reachable from the call graph.
21. The computer program product as described in claim 18 wherein the binary analysis generates the behavior model by micro-executing code reachable from an entrypoint of the container image.
22. The computer program product as described in claim 17 wherein the behavior model is instantiated in a hardware node in a Container Orchestration Engine (COE), and wherein the container image executes as a micro-service.
23. The computer program product as described in claim 22 , wherein the program code is further configured to enrich the behavior model based on information representing one or more valid system calls seen by one or more other containers running on the hardware node.
24. The computer program product as described in claim 17 , wherein the program code configured to determine whether the container image is executing in a manner inconsistent with the behavior model executes continuously and the automated action is one of: mitigation, notification, sandboxing and logging.