Embedded Container Control Plane for Clustered Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clustered storage systems face challenges such as increased costs and maintenance requirements due to resource overloading, data synchronization difficulties, and replication delays, particularly when managing distributed applications across multiple storage appliances.
Innovation Solution
Implementing an embedded container-based control plane with a central management database synchronously replicated to one peer appliance and asynchronously to others, allowing for distributed command processing and orchestration across primary and peer appliances, which improves scalability and reduces the need for data re-synchronization during appliance failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a separate management server and database are used to manage storage appliances, then centralized management is achieved, but system costs and maintenance requirements increase due to resource overloading
Solution Approach 1:
The patent divides the management function into segments: a centralized management server handles high-level orchestration while individual storage appliances maintain local management databases for operational decisions. This segmentation reduces the resource burden on the central server and enables distributed autonomy, resolving the contradiction between centralized management and system complexity.
Solution Approach 2:
Each storage appliance maintains a local copy of the management database, allowing it to operate autonomously without constant connection to the central server. This copying approach enables the system to function with reduced central server resources while maintaining management capabilities, addressing the cost and complexity issues.
2Reliability
If a clustered database is used across multiple storage appliances, then data consistency can be maintained, but system complexity increases and at least three storage appliances are required
Solution Approach 1:
The patent implements local quality by allowing each storage appliance to have its own management database with customized access rights and operational characteristics. This enables data consistency through local control rather than requiring a complex clustered database configuration with minimum three-appliance requirements.
Solution Approach 2:
The system dynamically adjusts database replication and synchronization based on operational needs rather than requiring a fixed clustered database configuration. This dynamic approach maintains data consistency while reducing system complexity and appliance requirements.
3Reliability
If a primary database with replicated secondary databases is used, then failover capability is provided, but replication lag causes delays in transaction propagation
Solution Approach 1:
The patent implements preliminary action by pre-configuring each storage appliance with a local management database that contains replicated data from the central database before failures occur. This ensures that when failover is needed, the local databases already contain the necessary information, eliminating replication lag delays and enabling immediate operational continuity.
Solution Approach 2:
The local management databases act as intermediaries between the central management server and the storage appliances' operational systems. This intermediary layer buffers the replication lag by maintaining local copies of data, allowing failover to occur without waiting for real-time replication propagation.
4Ease of operation
If storage management requests are routed to a central management server, then centralized control is achieved, but processing resources may outpace existing resources requiring costly upgrades
Solution Approach 1:
The patent segments the processing workload by having individual storage appliances handle their own management operations locally using their own databases, rather than routing all requests to a central server. This distribution eliminates the bottleneck of centralized processing resources and removes the need for costly upgrades to handle increased load.
Solution Approach 2:
Each storage appliance performs self-service management operations using its local database, independently handling its own configuration and operational tasks. This self-service approach eliminates the need for centralized processing resources and prevents resource overloading, allowing the system to scale without costly upgrades.
Data Source
AI summary
Techniques for providing an embedded container-based control plane for clustered environments. A clustered storage system includes one to many storage appliances, in which one storage appliance is designated as a “primary appliance” and any other storage appliance(s) is/are designated as a “peer appliance(s).” The primary appliance includes a central database, one peer appliance includes a synchronously replicated database, and any other peer appliances each include an asynchronously replicated database. The primary appliance further includes a global management IP for managing access to the central database. The primary and peer appliances each further include a control plane container for implementing command processing services. The control plane container of the primary appliance further implements command orchestration services. The clustered storage system with central and replicated databases and a distributed control plane avoids the need to re-synchronize data at the time of storage appliance failure and improves scalability.