Load-balancing for achieving transaction fault tolerance
AssigneeAmazon Technologies, Inc.
InventorDavid Walker
Life of this patent
Estimating the term…
Abstract
A system for servicing a request in a distributed computing environment includes a plurality of computing devices configurable to implement a first compute instance group. The first instance group includes a first plurality of compute instances of a provider network. One or more of the computing devices is configured to implement a first network device that is configured to receive a first service request from a client directed to an application and to provide the first service request to each of the first plurality of compute instances. The application is run on each of the first plurality of compute instances.
Claims15
A method of servicing a request in a distributed computing environment, comprising:
- configuring a plurality of computing devices to implement a plurality of compute instances of a provider network;
- allocating one or more of the computing devices to implement a network device that receives network traffic directed to an application;
- allocating each of the plurality of compute instances to execute the application;
- receiving, by the network device, a service request from a client directed to the application;
- providing, by the network device, the service request to each of the plurality of compute instances;
- receiving, by the network device, a service request response from each of the plurality of compute instances;
- in response to a majority of the service request responses being identical, providing, by the network device, one of the identical service request responses to the client;
- and
- in response to a first service request response being different from a majority of the service request responses, generating an alarm identifying a first of the plurality of compute instances from which the first service request response was received and terminating the first compute instance.
A system for servicing a request in a distributed computing environment, comprising:
- a plurality of computing devices, each computing device including a processor and memory, the plurality of computing devices configurable to implement a first compute instance group, the first compute instance group comprising a first plurality of compute instances of a provider network;
- and
- one or more of the computing devices configured to implement a first network device configured to:
- receive a first service request from a client directed to an application, the application being run on each of the first plurality of compute instances;
- provide the first service request to each of the first plurality of compute instances;
- receive a service request response from each of the first plurality of compute instances;
- and
- terminate a compute instance of the first plurality of compute instances whose service request response is different from the service request responses of a majority of the plurality of compute instances.
A computing system, comprising:
- one or more processors;
- and
- memory including instructions, that upon being executed by the one or more processors, cause the one or more processors to:
- implement a first plurality of compute instances;
- implement a first network device that receives network traffic directed to an application running on each of the first plurality of compute instances;
- receive a first service request from a client directed to the application;
- provide the first service request to each of the first plurality of compute instances;
- receive a service request response from each of the first plurality of compute instances;
- and
- terminate a compute instance of the first plurality of compute instances whose service request response is different from the service request responses of a majority of the plurality of compute instances.
Dependencies are inferred from the claim text.