Amazon Technologies, Inc.

Jeffrey Lloyd Baker

Avinash Jaisinghani

Prashant Jayaraman

Jaykumar Harish Gosar

Reto Kramer

Venkates Paramasivam Balakrishnan

Christopher Whitaker

Daniel Lee Osiecki

Nishanth Shankaran

Adam M Thomas

Danny Rabbani

Pete Peerapong Janes

Abhijeet Kumar

Madhusudan Govindarajan

Techniques are described that enable a user to upgrade a stack of resources by providing a template that reflects the desired end state of the stack. Upon receiving a new template, the system automatically detects changes that should be performed and determines the order in which the changes should be performed. The system can also detect if the desired changes to the stack are a result of direct modifications; because parameters to the resources have changed; or the changes are indirectly caused by other dependency/attribute changing. Additionally, the system determines whether the changes require creating new resources or if the changes can be made to the resources live. In case of replacement of resources, the system will first create the new resource, move that new resource into the stack and remove the old resource(s). In case of failures, the system ensures that the stack rolls back to the initial state.

Using a template to update a stack of resources

Techniques for executing commands associated with system stacks using parallel workflows are described herein. A dependency representation based at least in part on a stack description is created. The stack description describes stack resource instance and dependencies between those resource instances. The dependency representation is then analyzed to determine sub-workflows each of which may be executed in parallel. The sub-workflows may be altered based on one or more runtime interdependencies and the stack command may be executed by performing the sub-workflows in a determined order.

Parallel asynchronous stack operations

Growing community of inventors

Issaquah, WA, United States of America

Jeffrey Lloyd Baker