Storage Policy Management for Failure Prediction in Hybrid Clouds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a hybrid cloud storage environment, existing management systems fail to promptly notify users of operation failures, leading to potential issues due to delayed awareness of unsuccessful operations, and it is difficult to determine available functions from abstract policy-based management.
Innovation Solution
A management device that includes a processor and memory device to execute a program, utilizing policy tables, operation origin and destination tables, and state tables to identify potential failures by determining the normality of operations and their outcomes, thereby preventing future problems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automatic rollback is implemented, then operation recovery is improved, but notification delay to users persists
Solution Approach 1:
The system performs preliminary actions by maintaining state information about operation origins and destinations before failures occur. The state table pre-stores operational states, enabling the system to quickly determine failure causes and notify users without delay when failures happen, while also enabling automatic rollback.
2Ease of operation
If policy-based management is used, then operation abstraction is improved, but determination of available functions becomes difficult
Solution Approach 1:
The state table acts as an intermediary between the abstract policy-based management layer and the concrete operational layer. It stores and manages state information about operation origins and destinations, enabling the system to determine available functions and failure causes without exposing the complexity of the underlying policies to users.
3Measurement precision
If comprehensive state tracking is implemented, then failure identification is improved, but system complexity increases
Solution Approach 1:
The system segments state information into distinct categories: operation origin states, operation destination states, and policy combinations. The state table is divided into multiple fields that separately track different aspects of operational state, enabling precise failure identification without creating an undifferentiated complex structure.
Data Source
AI summary
A management device manages one or more platforms, each having a storage device including one or more volumes, and stores: a policy table defining a policy concerning a volume as a resource and a copy policy concerning copy to define an operation of copying data to the resource; an operation origin table defining an operation origin in the platform where the operation is performed; and a state table for maintaining states of the operation origin. The processor extracts a combination of the resource and the operation from a combination of the policy concerning the volume and the policy concerning the copy. The processor identifies the operation origin from the combination of the resource and operation and determines whether the operation origin is normal, and determines whether the combination of the policies concerning the volume and the copy causes a failure on the operation origin determined to be abnormal.