Islanded Data Center Application Control Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
When a data center (DC) becomes unavailable to external computing entities due to loss of network connection or other disasters, the applications executing at the DC may continue to run without control, leading to potential unwanted operations and hindering effective failover to another DC.
Innovation Solution
A method is implemented at the DC to determine when applications are no longer controllable by external entities, triggering a disaster recovery mode. This involves communicating an indication to applications already executing and those starting to execute, allowing them to operate in disaster recovery mode, which may include ceasing functions or shutting down to prevent unwanted operations and facilitate failover.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If applications continue to execute at the DC after loss of external control, then application availability is maintained, but unwanted operations may occur and failover is hindered
Solution Approach 1:
The system performs preliminary actions by detecting the loss of external control before it fully takes effect. The monitoring mechanism identifies when the DC is no longer controllable by external computing entities and triggers the disaster recovery mode initiation, allowing applications to be guided to cease functions or shut down proactively rather than reactively after harmful operations have occurred.
Solution Approach 2:
The system implements feedback through continuous monitoring of external control connectivity. The monitoring mechanism provides real-time feedback on the control status to the applications, enabling them to adjust their operation based on the current state. When loss of control is detected, the feedback loop activates the disaster recovery mode communication to applications, ensuring they respond appropriately to prevent unwanted operations.
2Object-affected harmful factors
If applications are stopped or shut down to prevent unwanted operations, then harmful effects are reduced, but service continuity is interrupted
Solution Approach 1:
The system applies dynamics by making the application shutdown process configurable and adaptable. The disaster recovery mode can be configured with different shutdown policies depending on the specific scenario. Applications can be stopped, shut down, or configured to cease specific functions based on the severity and type of loss of control, allowing flexible response that balances preventing harmful operations with maintaining service continuity where possible.
3Reliability
If failover is implemented to another DC, then application control is restored, but system complexity increases
Solution Approach 1:
The system implements self-service by enabling applications to autonomously detect and respond to loss of external control. The disaster recovery mode is automatically initiated and communicated to applications without requiring constant manual intervention or complex centralized control. Applications self-manage their shutdown or function cessation based on the detected condition, reducing the operational complexity of failover while maintaining reliability.
Data Source
AI summary
A data center, DC includes applications that are controllable by one or more computing entities external to the DC. However, when the DC becomes unavailable to the computing entities, the applications are no longer controllable by the computing entities, which is undesirable. Disclosed techniques include the DC determining, based on an indication that the apps are no longer controllable by the computing entities, that the applications are to operate in a disaster recovery mode. Disclosed techniques include causing an indication that the applications at the DC are to operate in the disaster recovery mode to be communicated to applications that are already executing at the DC when the determination is made and to applications that start to execute at the DC after the determination is made. Disclosed techniques also include a further DC providing a failover for the DC.