Distributed Orchestration Engine Cloud-Agnostic Deployment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional cloud orchestration techniques are inefficient, create bottlenecks, and decrease fault tolerance due to their cloud-specific nature, requiring duplicate effort and higher costs for maintaining and debugging orchestration code across multiple clouds, and are susceptible to single points of failure when using a single node for orchestration.
Innovation Solution
A Distributed Orchestration Engine (DOE) facilitates cloud-agnostic orchestration and deployment across multiple clouds using a shared event message bus with persistence, allowing nodes to obtain an ordered set of events based on execution sequences and select events ready for execution, independent of specific cloud implementations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional cloud-specific orchestration techniques are used, then orchestration can be performed for specific cloud platforms, but it requires duplicate effort and higher costs for maintaining and debugging code across multiple clouds
Solution Approach 1:
The patent implements a universal orchestration engine that can operate across multiple cloud platforms (AWS, Azure, GCP, etc.) through a single codebase. The system uses cloud-agnostic event sources and event handlers that can be configured for different cloud providers without requiring duplicate orchestration code, thereby achieving multi-functionality and reducing maintenance effort.
Solution Approach 2:
The patent employs event source copying where event definitions and handlers are replicated across different cloud platforms through configuration rather than code duplication. The same event processing logic is copied and applied to different cloud event sources, allowing the system to handle multiple cloud platforms while maintaining a single source of truth for orchestration logic.
2Device complexity
If a single node is used for orchestration, then the system structure is simple, but it creates bottlenecks and decreases fault tolerance
Solution Approach 1:
The patent segments the orchestration system into multiple independent nodes that can operate in parallel. Each node can process events independently, eliminating the bottleneck of a single node. The system divides the orchestration workload across multiple instances, improving both throughput and fault tolerance through distributed processing.
Solution Approach 2:
The patent changes the parameter of node quantity from single to multiple, transforming the system from a monolithic orchestration point to a distributed network of nodes. This parameter change enables the system to scale horizontally, improving reliability and fault tolerance while maintaining manageable complexity through standardized node interfaces.
3Productivity
If cloud-specific orchestration code is used, then optimization for specific cloud platforms is achieved, but it increases costs and requires duplicate maintenance effort
Solution Approach 1:
The patent introduces an intermediary layer consisting of standardized event sources and event handlers that mediate between cloud-specific event triggers and cloud-agnostic orchestration logic. This intermediary architecture allows the system to maintain high productivity through optimized cloud-specific event handling while reducing maintenance time by keeping the core orchestration logic universal and independent of specific cloud providers.
Data Source
AI summary
Embodiments disclosed facilitate distributed orchestration and deployment of a cloud based distributed computing application. In some embodiments, the distributed orchestration may be based on a cloud agnostic representation of the application. In some embodiments, an ordered set of events may be obtained based on a starting sequence for the plurality of components, where each event in the ordered set is associated with a corresponding set of prerequisites for initiating execution of the event. Event identifiers corresponding to the ordered set of events may be placed on an event message bus with persistence that is shared between a plurality of distributed nodes associated with a cloud. Upon receiving an indication of completion of prerequisites for one or more events, a plurality of nodes of a distributed orchestration engine may select the one or more events corresponding to the one or more selected event identifiers for execution on the cloud.