Amazon Technologies, Inc.

Nathan Lee Burns

Bin Wang

Nicholas Patrick Wilt

Ashutosh Tambe

Scott Wright Heath

Kevin Cheek

Michelle Lynn Dobbs

Prabu Ramachandran

Nafea Bshara

Collin Charles Davis

Nelamangal Krishnaswamy Srinivas

Pavan Kumar Surishetty

Austin Garbelman

Christopher Michael VanderKnyff

Matthew Sterling Leese

Siddhartha Shankara Rao

Thaddeus C Pritchett

John R Garvey

Richard Weatherly

Jesen Kwok Ha

Andrew Wolfe

Ben Linville

Ismael Gonzales

Michael Andrew Gibson

Methods, systems, and computer-readable media for virtualizing graphics processing in a provider network are disclosed. A virtual compute instance is provisioned from a provider network. The provider network comprises a plurality of computing devices configured to implement a plurality of virtual compute instances with multi-tenancy. A virtual GPU is attached to the virtual compute instance. The virtual GPU is implemented using a physical GPU, and the physical GPU is accessible to the virtual compute instance over a network. An application is executed using the virtual GPU on the virtual compute instance. Executing the application generates virtual GPU output that is provided to a client device.

Virtualizing graphics processing in a provider network

A Data Delivery Service (DDS) is described, which is a service in a multi-tenant environment that transmits satellite data between a satellite antenna and a user instance. The DDS transports the antenna data to a different region, which allows a user to reuse their infrastructure for multiple antenna sites, thereby, reducing their infrastructure footprint and costs. Gateway instances can be launched at scheduled times in different regions and a secure communication channel can be established between the gateway instances to establish inter-region communication.

Service connecting antennas to remote regions

Satellites may be integrated into a provider network for cloud-based services by utilizing a satellite-deployable computing device to execute one or more cloud services that are accessible by client devices of the provider network via respective associated Application Programming Interfaces (APIs). In some embodiments, the satellite-deployable computing device may process payload data of the satellite in accordance with one or more cloud-based services indicated in control instructions originating from a client device of the provider network. In some embodiments, the processed data may be transmitted to a ground station in accordance with one or more cloud-based services.

Satellite transmission module for ground station

A matchmaking service can match a satellite owner/operator with antennas that are available and capable of servicing a communication request from the satellite owner/operator (e.g., to exchange data with the satellite owner/operator's satellite). The communication request may be received at a ground station service, and information included therein may be compared to capabilities of antennas registered with the ground station service to determine antennas that are compatible with the request. The ground station service may reserve a time slot for a selected compatible antenna to service the request and control the selected compatible antenna (e.g., to track and communicate with the satellite) during the time slot. In some embodiments, the matchmaking of satellite owners/operators to antennas uses models to determine communication parameters for a communication request and to determine capabilities of antennas registered with the ground station.

Antenna orchestration as a service

Systems and methods are provided for dynamically routing packets using multi-flow and multi-path multiplexing connections. A first computing device and second computing device communicate via a plurality of data flows, which may be routed across various network paths. Each flow is defined by a set of network addresses, a set of ports, and a protocol specification, such as UDP. The second device sends information to the first device regarding the various data flows. The first device may send probing packets to facilitate collecting data flow information. Information may include latency, packet loss, and other values. Based on the information received, the first device may select or prioritize data flows to mitigate congestion, and address performance criteria. The first device may also transmit information regarding selected data flows to other devices, allowing the other devices to utilize a selected data flow even if they lack multiplexing capability.

Multi-path routing

Systems and methods are described for a connection service that identifies connections for providing a data flow between a client computing device and a target computing device. A client computing device can send an API-based connection request to various servers hosting the connection service. The connection service processes the requests to determine a routing path for the connection. Advantageously, the connection service determines various routing paths and further identifies a connection for the data stream using that routing path. Finally, a control service provides control and relays requests to initiate data flows for resources configured for a data flow (e.g., a remote desktop session).

Connection service with network routing

Methods, systems, and computer-readable media for application-specific virtualized graphics processing are disclosed. A virtual compute instance is provisioned from a provider network. The provider network comprises a plurality of computing devices configured to implement a plurality of virtual compute instances with multi-tenancy. A virtual GPU is attached to the virtual compute instance. The virtual GPU is selected based at least in part on requirements of an application. The virtual GPU is implemented using a physical GPU, and the physical GPU is accessible to the virtual compute instance over a network. The application is executed using the virtual GPU on the virtual compute instance.

Application-specific virtualized graphics processing

Systems and methods are described for a connection service that identifies routing paths for providing a data flow between a client computing device and a target computing device. A client computing device can send an API-based connection request to various servers hosting the connection service. The connection service processes the requests to determine a routing path for the connection. Advantageously, the connection service selects a routing path from a plurality of available routing paths if a client computing device is associated with a network address translator. The connection service may provide an externally addressable network address if it determines that no such address is available for routing of a data flow.

Systems and methods are described for a routing service that identifies routes for providing a resource with the least cost for providing a resource configured for a data flow. The routing service can operate on various servers with varying functionalities in a distributed network topology. After establishing an application-programming interface ('API') session, the routing service can send API-based requests to the various servers to obtain API-based routing information. The routing service can use the routing information to determine that a network performance metric exceeds a threshold for providing a resource on a requested route. Advantageously, the routing service can determine alternative routes using the API routing information and, further, identify the least cost route of those alternative routes. Finally, the routing service can provide a resource configured for a data flow (e.g., a video conference) via the least cost route.

Network routing utilization of application programming interfaces

A computing device can be utilized to submit content to a collaboration service for posting on a collaboration channel. In response to receiving a posting request, the collaboration service identifies a context associated with the request. The context might include the location of the computing device submitting the request, the authenticated identity of the user submitting the request, or access rights associated with the content to be posted. The collaboration service can utilize the context to identify one or more collaboration channels to which the content can be posted. The user might be permitted to select one or more of the identified channels. Once the content has been posted to the selected channel, or channels, the collaboration service can utilize the context to perform other actions such as, but not limited to, initiating a communications session between authorized users of the collaboration channel to which the content was posted.

Growing community of inventors

Mountain View, CA, United States of America

Nathan Lee Burns