Application-Specific Disaster Recovery Policies for Edge-to-Cloud Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing disaster recovery solutions are not tailored to the specific needs of edge sites, which often operate in harsh environments and handle diverse workloads, leading to inefficiencies and increased failure risks.
Innovation Solution
A workload-aware disaster recovery system that dynamically creates application-specific DR policies, allowing for preferential failover and failback responses to DR events by utilizing a DR smart agent that monitors workload patterns and adheres to recovery time and point objectives, and a data replication fabric for seamless edge-to-cloud data movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing disaster recovery solutions are used at edge sites, then basic failover capability is provided, but the solutions are not tailored to specific edge needs leading to inefficiencies and increased failure risks
Solution Approach 1:
The system segments disaster recovery policies by application type, creating specialized policies for different workload categories (e.g., stateless applications, stateful applications, database applications). Each segment receives tailored recovery instructions based on its specific requirements, enabling edge sites to handle diverse workloads appropriately while maintaining overall system reliability.
Solution Approach 2:
The patent implements local quality by providing customized DR policies for different applications at edge sites rather than a uniform approach. Each application receives policy parameters optimized for its characteristics (e.g., recovery time objectives, data retention requirements), ensuring that the DR solution adapts to local edge site conditions and specific application needs.
2Ease of manufacture
If a one-size-fits-all disaster recovery approach is used, then implementation simplicity is maintained, but efficiency decreases and failure risks increase for diverse edge workloads
Solution Approach 1:
The system achieves universality by creating a framework that handles multiple application types through a common policy structure. The classification mechanism universally applies to various workloads (stateless, stateful, database applications) while automatically adjusting policy parameters, thus maintaining implementation simplicity across diverse edge site scenarios without sacrificing efficiency.
Solution Approach 2:
The patent utilizes parameter changes by adjusting specific policy parameters (recovery time objectives, data retention periods, failover thresholds) based on application classification. This allows the system to maintain a universal policy framework while dynamically modifying parameters to optimize efficiency for each application type, resolving the contradiction between simplicity and effectiveness.
Data Source
AI summary
Example implementations relate to application-specific policies for failing over from an edge site to a cloud. When an application becomes operational within an edge site, a discovery phase is performed by a local disaster recovery (DR) agent. I/O associated with a workload of the application is monitored. An I/O rate for data replication that satisfies latency characteristics of the application is predicted based on the incoming I/O. Based on results of tests against multiple clouds indicative of their respective RTO/RPO values, information regarding a selected cloud to serve as a secondary system is stored in an application-specific policy. The application-specific policy is transferred to a remote DR agent running in the selected cloud. Responsive to a failover event, infrastructure within a virtualized environment of the selected cloud is enabled to support a failover workload for the application based on the application-specific policy.