Workflow Action Graph Translation for Cloud Resource Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems that manage workflow actions in cloud platforms require significant computing resources due to the creation and maintenance of large graphs of nodes for each action, leading to inefficiencies and resource wastage, especially when dealing with repeated actions and multiple clients.
Innovation Solution
The system translates a set of workflow actions into a graph with a constant number of nodes, independent of the number of actions, using a sequence tracking parameter to maintain state and execute nodes only when actions are completed, thereby reducing memory and processing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large graph of nodes is created for each workflow action to track state, then workflow execution accuracy is improved, but memory usage and processing resources increase significantly
Solution Approach 1:
The patent extracts the state tracking functionality from the graph structure itself and relocates it to external storage (database or file system). Instead of maintaining state within the in-memory graph, the system stores state data separately, allowing the graph to remain small and memory-efficient while still accurately tracking workflow state through external persistence.
Solution Approach 2:
The patent creates a single reusable graph template that can serve multiple workflow actions and clients. Instead of creating separate large graphs for each action, one universal graph structure is designed that can be instantiated and configured for different workflows, reducing overall memory consumption while maintaining execution accuracy across multiple scenarios.
2Reliability
If a large graph of nodes is created for each workflow action, then complete state tracking is achieved, but processing time and computational resources increase
Solution Approach 1:
The patent extracts computationally intensive state tracking operations from the in-memory graph processing and moves them to external storage systems. State transitions and workflow progress are recorded in databases or file systems, reducing the processing burden on the graph engine and enabling faster execution without sacrificing tracking completeness.
Solution Approach 2:
The patent pre-defines the graph structure and node templates before workflow execution begins. By preparing the graph framework in advance with all necessary nodes and relationships defined, the system avoids dynamic graph construction during execution, significantly reducing processing time while maintaining complete state tracking capabilities.
3Reliability
If separate graphs are created for each client and action, then client-specific workflow accuracy is improved, but resource wastage increases due to redundancy
Solution Approach 1:
The patent designs a universal graph template that can be shared across multiple clients and workflow actions. This single graph structure incorporates parameter placeholders and configuration options that allow it to be adapted to different client requirements without creating separate graphs, eliminating redundancy while preserving client-specific workflow accuracy through configurable parameters.
Solution Approach 2:
Instead of creating entirely separate graphs for each client, the patent uses copying of the universal graph template with parameter substitution. The base graph structure is copied and configured with client-specific parameters, maintaining accuracy for each client while sharing the common structural framework, thus reducing resource consumption compared to creating independent graphs from scratch for each client.
Data Source
AI summary
A database system translates a received set of workflow actions into a graph of nodes. The graph of nodes includes a node set corresponding to each node type of the set of workflow actions. Each node set may be re-executable for each instance of an action type of the set of workflow actions. Upon completion of an action, the graph of nodes is executed. Execution of the graph of nodes includes executing processor executable instructions corresponding to a subset of the set of workflow actions. The subset may be determined based on the completed action. The completed action may be determined based on a sequence tracking parameter.