Intelligent Application Clusters Disk Arbitration for Storage Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage area networks face limitations in providing high availability and manageable services due to the lack of effective mechanisms for ensuring data integrity and failover capabilities, especially in scenarios involving network partitions and hardware or network failures.
Innovation Solution
The implementation of intelligent application clusters with disk arbitration mechanisms, a virtual logical unit number server (VS) architecture that includes a resource manager, arbitration module, heartbeat module, and state manager to ensure data integrity and facilitate failover, using protocols like VRRP for heartbeat management and IP address takeover, and synchronization of configuration information between active and standby VS.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple virtual logical unit number servers (VS) are deployed in a service-group cluster to provide redundancy, then high availability is improved, but data integrity is compromised when network partitions occur and multiple VS instances simultaneously write to shared regions
Solution Approach 1:
An arbitration module is introduced as an intermediary between multiple VS instances and shared storage resources. This module receives write requests, determines which VS instance has arbitration rights based on arbitration information (such as VS instance identifiers), and selectively permits or blocks write operations. This mediator mechanism prevents simultaneous writes from multiple VS instances during network partitions, thereby eliminating data corruption while preserving high availability through proper failover.
Solution Approach 2:
The system implements a feedback mechanism where VS instances continuously exchange heartbeat signals and arbitration information with each other and with the arbitration module. The arbitration module monitors the operational status of VS instances and dynamically adjusts write permission grants based on current cluster state. This feedback loop ensures that only the active master VS can write to shared regions at any given time, preventing data corruption while maintaining service continuity.
2Productivity
If configuration changes are made at an active VS, then service responsiveness is improved, but configuration consistency deteriorates when the standby VS is not properly synchronized
Solution Approach 1:
A configuration synchronization mechanism with feedback is implemented where the active VS sends configuration change notifications to the standby VS through TCP connections. The standby VS acknowledges receipt and verification of configuration updates, providing feedback to the active VS. This ensures that configuration information remains consistent across all VS instances while allowing the active VS to respond immediately to service requests without waiting for synchronization completion.
3Reliability
If disk arbitration mechanisms are implemented to ensure only one VS controls a disk-group, then data integrity is improved, but system complexity increases due to additional arbitration protocols and modules
Solution Approach 1:
The arbitration module is designed as a multi-functional component that combines multiple responsibilities: monitoring VS instance status through heartbeats, managing arbitration rights based on VS identifiers, controlling write permissions to shared storage, and coordinating failover operations. By consolidating these functions into a single universal module, the system achieves data integrity through comprehensive arbitration while minimizing the increase in overall system complexity compared to having separate mechanisms for each function.
Data Source
AI summary
Methods and apparatus provide formation and management of intelligent application clusters in a storage area network. Disk arbitration mechanisms ensure that a cluster is owned by a single member. In the event of a network partition, each cluster member involved arbitrates to gain ownership of a cluster. High availability mechanisms allow monitoring of system resources and effective failover capabilities.