Incremental Database Recovery for Cross-Region Shard Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database systems face challenges in providing full-region disaster recovery while minimizing resource consumption, especially in scenarios where regional disasters occur, due to the high costs and inefficiencies of reserving duplicate resources in secondary regions.
Innovation Solution
A system that logically splits tenant data into shards distributed across storage nodes, with backups stored in an object storage service, allowing for incremental recovery of metadata and shards on demand, distributing workload across multiple regions, and utilizing a margin of free resources in each region to minimize resource reservation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full database resources are reserved and operated in secondary regions for disaster recovery, then reliability is improved, but device complexity and resource consumption increase
Solution Approach 1:
The patent segments the database into shards that are distributed across storage nodes. During disaster recovery, only the necessary shards need to be recovered and redistributed across available storage nodes in the affected region, rather than recovering entire database instances. This segmentation enables partial recovery and reduces the complexity of full system restoration.
Solution Approach 2:
The patent implements partial replication where only critical or frequently accessed database shards are replicated to secondary regions, rather than maintaining complete copies of all database resources. This partial action approach provides sufficient disaster recovery capability while significantly reducing the resource reservation burden in secondary regions.
2Use of energy by moving object
If asynchronous replication is used to reduce latency costs, then resource consumption is reduced, but data consistency and reliability deteriorate
Solution Approach 1:
The patent performs preliminary backup of database shards to remote storage locations before disasters occur. These pre-positioned backups are geographically distributed and can be quickly restored to storage nodes in the event of a regional disaster, providing reliability without requiring continuous synchronous replication and its associated resource consumption.
Solution Approach 2:
The patent creates copy shards that are redistributed across storage nodes during normal operations. These copies serve as ready-to-use backups that can be rapidly deployed during disaster recovery, eliminating the need for continuous data synchronization while maintaining data availability and consistency.
3Use of energy by moving object
If manual activation of replica database instances is implemented, then resource consumption is reduced, but productivity and recovery speed worsen
Solution Approach 1:
The patent implements automated disaster recovery orchestration that detects regional failures and triggers shard redistribution and recovery processes without manual intervention. The system automatically identifies affected shards, selects appropriate storage nodes for recovery, and executes the restoration process, thereby maintaining fast recovery speed while avoiding the resource overhead of manually managed replica instances.
Solution Approach 2:
The patent employs dynamic shard redistribution where the system automatically adapts to failure conditions by reallocating shards across available storage nodes. This dynamic response enables rapid recovery by leveraging existing distributed shard copies and available infrastructure, eliminating recovery delays while maintaining resource efficiency.
Data Source
AI summary
Systems and methods include storage of shards of first database tables of a first tenant in a first plurality of storage nodes located in a first region, each shard associated with a first database table and a key range of the first database table, storage of shards of second database tables of a second tenant in a second plurality of storage nodes located in a second region, each shard associated with a second database table and a key range of the second database table, storage of backups of the shards of the first database tables of the first tenant in a plurality of backup locations located in a region different from the first region, and recovery of the backups of the shards of the first database tables of the first tenant from the backup layer to the second plurality of storage nodes.