Non-monotonic Convergence for Desired State Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Desired State Configuration (DSC) systems face challenges in configuring a target system for a desired state when the current state of a service needs to be replaced, as stopping the service moves away from the desired state, leading to configuration failures.
Innovation Solution
The implementation of non-monotonic eventual convergence (NMEC) allows a DSC system to configure a target system by temporarily moving further away from the desired state if necessary, enabling the replacement of services and eventually reaching the desired operational state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a service is stopped to replace its executable file, then the file replacement can be completed, but the service moves away from the desired operational state
Solution Approach 1:
The system performs preliminary actions by stopping the service temporarily before replacing the executable file, then restores the service to its operational state. This preliminary interruption allows the file replacement to proceed while the system plans to return to the desired state, resolving the contradiction between file replacement capability and service operational reliability.
Solution Approach 2:
The system inverts the conventional approach by allowing temporary deviation from the desired operational state as an acceptable intermediate step. Instead of requiring the service to remain operational throughout the process, the system accepts a temporary stoppage, performs the file replacement, and then restores operation, thus solving the contradiction by reversing the expected sequence of maintaining state versus performing replacement.
2Device complexity
If conventional DSC systems require monotonic convergence toward desired state, then configuration simplicity is maintained, but updates fail when temporary deviation is needed
Solution Approach 1:
The system introduces dynamics to the configuration process by allowing temporary deviations from the desired state when necessary for updates. Instead of enforcing a rigid monotonic convergence, the system dynamically adjusts by permitting service stoppage during file replacement, then restoring the operational state, thereby improving update success rates while maintaining relative configuration simplicity.
Solution Approach 2:
The system changes the parameter of state convergence from strictly monotonic to non-monotonic, allowing temporary movement away from the desired state. This parameter change enables the service to be stopped temporarily during file replacement, then restored, thus improving update reliability without significantly increasing configuration complexity.
3Duration of action of stationary object
If a service remains running during file replacement, then service continuity is maintained, but file replacement cannot be performed
Solution Approach 1:
The system performs preliminary action by stopping the service temporarily to enable file replacement, then restores the service to maintain continuity. This preliminary interruption is necessary to allow the file replacement operation, and the subsequent restoration ensures service continuity is reestablished, resolving the contradiction between maintaining continuity and enabling replacement.
Solution Approach 2:
The system introduces an intermediary state where the service is temporarily stopped between the initial operational state and the final restored operational state. This intermediary stoppage enables the file replacement to occur, and the service is then restored to its running state, thus mediating between the conflicting requirements of service continuity and file replacement capability.
Data Source
AI summary
Technologies are provided for non-monotonic eventual convergence for desired state configuration (“DSC”). In some situations DSC cannot move toward a desired state without first moving further from that state. For example, a file that needs to be replaced with a newer version but that is currently executing (desired state of “operating”) cannot be replaced without first being stopped. But stopping moves in the wrong direction relative to the desired state, which is to have the service operating. This moving away so as to be able to move forward is a problem for conventional DSC systems that results in failures. The solution to this problem is non-monotonic eventual convergence (“NMEC”) which enables a DSC system to configure a target system for a desired state by moving further away from a desired state if that is needed to eventually reach the desired state.