Privacy preserving introspection for trusted execution environments
AssigneeRed Hat, Inc.
InventorsMichael Bursell, Michael Tsirkin
Life of this patent
Estimating the term…
Abstract
A system includes a memory, a processor in communication with the memory, a hypervisor, and a trusted execution environment (TEE). The TEE is provisioned with a workload and includes an introspection module. The introspection module is configured to execute an introspection command according to an introspection policy. The introspection command is configured to validate at least one memory access associated with the workload. The introspection module is also configured to determine a status of a result of the introspection commands, wherein the status is one of a failure status and a success status.
Claims19
A system comprising:
- a memory;
- a processor in communication with the memory;
- a hypervisor;
- and
- a trusted execution environment (TEE), wherein the TEE is provisioned with a workload and includes a runtime executing the workload, wherein the runtime is configured to:
- execute an introspection command according to an introspection policy, wherein the introspection policy identifies connectors and accelerators available to the introspection command, wherein the connectors can connect to external databases for the introspection command, wherein the accelerators are configurable to convert resulting information of the introspection command into one or more different forms and to determine requisite connector rules to build a connector, the introspection command configured to validate at least one memory access associated with the workload, and
- based on a status of a result of the introspection command, wherein the status is one of a failure status and a success status, implement a security enforcement mechanism that comprises at least one of pausing the workload, shutting down the TEE, or stopping execution of the TEE.
A method comprising:
- provisioning a trusted execution environment (TEE) with a workload for execution within a runtime, wherein the workload can comprise an introspection module;
- executing, within the runtime, an introspection command according to an introspection policy, wherein the introspection policy identifies connectors and accelerators available to the introspection command, wherein the connectors can connect to external databases for the introspection command, wherein the accelerators are configurable to convert resulting information of the introspection command into one or more different forms and to determine requisite connector rules to build a connector, the introspection command configured to validate at least one memory access associated with the workload;
- and
- based on a status of a result of the introspection command, wherein the status is one of a failure status and a success status, implementing a security enforcement mechanism that comprises at least one of pausing the workload, shutting down the TEE, stopping execution of the TEE.
A system comprising:
- a memory;
- a processor in communication with the memory;
- a trusted execution environment (TEE), wherein the TEE is provisioned with a workload, wherein the workload can comprise an introspection module;
- and
- a runtime executing the workload within the TEE, wherein the runtime is configured to:
- load a pre-generated memory accessing code into the TEE;
- receive introspection commands through an introspection channel, wherein the introspection channel is created between the runtime and an owner, wherein the introspection policy identifies connectors and accelerators available to the introspection command, wherein the connectors can connect to external databases for the introspection command, wherein the accelerators are configurable to convert resulting information of the introspection command into one or more different forms and to determine requisite connector rules to build a connector;
- execute the introspection commands according to an introspection policy, wherein the introspection policy is enforced by the runtime, wherein the introspection commands are configured to validate a memory access in the pre-generated code;
- and
- based on a status of a result of the introspection commands, wherein the status is one of a failure status and a success status, implement a security enforcement mechanism that comprises at least one of, pausing the workload, shutting down the TEE, stopping execution of the TEE.
Dependencies are inferred from the claim text.