Parallel Multi-Region Workflow Execution With Local Status Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional relational database management systems (RDBMS) are limited to single-region deployments and require extensive reconfigurations for multi-regional configurations, necessitating adaptations for non-sequential Globally Unique Identifiers (GUIDs) and database referential integrity to support multi-regional multi-master datastores.
Innovation Solution
Implementing a system that executes cloud-based processing requests in parallel across multiple regional servers, utilizing local instances of a status database to replicate modifications and synchronize job statuses, thereby facilitating parallel execution without reliance on external data stores prone to communication interruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional RDBMS are used for single-region deployment, then system simplicity is maintained, but multi-regional capability is lost
Solution Approach 1:
The patent segments the monolithic RDBMS into multiple independent regional instances, each operating autonomously in its own region. Each region maintains its own database instance with local GUID generation and referential integrity enforcement, eliminating the need for complex cross-region configuration while achieving multi-regional capability.
Solution Approach 2:
The patent introduces an intermediary mechanism for cross-region operations that handles GUID validation and referential integrity checks without requiring direct coupling between regional instances. This intermediary layer enables multi-regional functionality while preserving the simplicity of individual region configurations.
2Productivity
If parallel job execution is implemented across multiple regions, then processing speed is improved, but coordination complexity increases
Solution Approach 1:
The patent merges the job status tracking functionality into the existing regional database instances, eliminating the need for separate coordination systems. Each region's database instance independently tracks its own job status while maintaining compatibility with cross-region job definitions, simplifying the overall coordination architecture.
Solution Approach 2:
Each regional database instance autonomously manages its own job execution and status tracking without requiring external coordination services. The instances self-serve by locally enforcing GUID validity and referential integrity, reducing the complexity of inter-region job coordination while maintaining parallel execution capabilities.
3Reliability
If external data stores are used for job status reporting, then centralized monitoring is achieved, but communication interruptions cause delays
Solution Approach 1:
The patent implements local quality by enabling each regional database instance to independently store and provide job status information locally. This eliminates the single point of failure represented by external data stores, ensuring that job status availability is maintained even when communication between regions is interrupted.
Data Source
AI summary
Techniques for performing parallel processing jobs across regionally-distributed servers may include executing a requested job in parallel across two or more regional cloud-based servers, the status of parallel jobs may be locally monitored using respective local instances of a status-tracking database. The multiple instances may replicate modifications of one instance of the status database to other instances of the status database to synchronize the status of regional jobs. Replicating modifications across the respective regional instances of the status database facilitates the parallel executing of the jobs across regions while reducing or avoiding unrecoverable delays if the tasks/job hangs or otherwise fails.