公开(公告)号：US11614957B1

公开(公告)日：2023-03-28

申请号：US16869376

申请日：2020-05-07

Applicant: Amazon Technologies, Inc.

Inventor： Vijay Dheeraj Reddy Mandadi ,  Raviprasad V Mummidi

IPC: G06F9/455 ,  G06F9/50

Abstract: Systems and methods are provided for on-demand code execution that uses a client's host computing environment with native hypervisors. The client's host computing device receives a configuration application library. An agent process is executed on the host computing device from the configuration application library. The agent process is executed in association with a first user profile that gives the process rights to configure the client host computing device. The agent process receives, from a service external to the client host computing environment/device/native hypervisor, a code-execution request on behalf of a second user profile. The agent process provisions an execution environment on behalf of the second user profile. The execution environment runs on the native hypervisor. The code instructions are executed in the execution environment under the second user profile. The agent process monitors the code execution and receives a status regarding the execution of the code in the execution environment. The agent process further transmits the status back to the service.

公开(公告)号：US12099923B1

公开(公告)日：2024-09-24

申请号：US16685188

申请日：2019-11-15

Applicant: Amazon Technologies, Inc.

Inventor： Vijay Dheeraj Reddy Mandadi ,  Raviprasad V Mummidi

IPC: G06N3/08 ,  G06F9/455 ,  G06N3/04

CPC classification number: G06N3/08 ,  G06F9/45533 ,  G06N3/04

Abstract: Applications may execute using resources from multiple environments or ecosystems, such as may include virtual resources from a virtual resource environment hosted on physical resources of a cloud provider environment. An event manager in the cloud provider environment can obtain virtualization performance data from the virtual resource environment, as well as performance data from within the cloud provider environment. This data can be fed to an inference engine that can correlate information from the separate environments, and these correlations can be used to generate recommendations for performance adjustments in either the physical or virtual resource environment.