Networked Storage Failover Without SCSI-3 Reservations

The patent extracts and eliminates the SCSI-3 reservation mechanism from the failover process. Instead of relying on reservations to prevent access, the system uses a simplified approach where the surviving node directly assumes control of storage devices without requiring reservation requests, thereby removing the source of delay while maintaining protection through alternative means

The patent implements preliminary actions by pre-configuring node identifiers and storage mappings before failure occurs. When failure happens, the surviving node already has the necessary information to immediately assume control without needing to execute reservation protocols, enabling faster failover while maintaining data protection

The patent segments storage access rights by identifying specific storage devices or partitions that can be independently controlled. This allows the failing node to be isolated from specific storage segments while other nodes retain write access to other segments, maintaining both control and versatility without requiring system-wide reservations

The patent applies local quality by differentiating access rights at the storage device level rather than system-wide. Each storage device or logical unit can have its own access control state, allowing selective access for different nodes depending on their operational status, thereby maintaining multi-node write capability where appropriate