Distributed Object Timestamps During Storage Link Disruptions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face inefficiencies in managing data across multiple storage arrays, particularly in coordinating data operations and maintaining data integrity during power failures or controller failures, leading to potential data loss and increased latency.
Innovation Solution
Implementing a storage system architecture with dual storage array controllers that offload device management responsibilities, utilize non-volatile RAM for buffering, and synchronize object creation times across multiple storage systems to ensure data consistency and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If synchronous negotiation of object creation time is implemented across multiple storage systems, then data consistency and reliability are improved, but system complexity and coordination overhead increase
Solution Approach 1:
A timestamp authority is introduced as an intermediary component that centralizes the generation and distribution of monotonically increasing timestamps. This mediator receives write requests from multiple storage systems, assigns sequential timestamps, and returns them to the originating systems. By concentrating timestamp management in a single authoritative source, the system achieves data consistency across distributed storage without requiring complex peer-to-peer synchronization protocols, thus improving reliability while controlling system complexity.
2Reliability
If dual storage array controllers are used to offload device management responsibilities, then system reliability and fault tolerance are improved, but device complexity and coordination requirements increase
Solution Approach 1:
The storage system is divided into two independent controllers, each managing specific device responsibilities. This segmentation allows the system to distribute device management tasks across multiple controllers, enabling fault tolerance since the failure of one controller does not compromise the entire system. Each controller can operate independently to handle read/write requests, and the system maintains data integrity through coordinated operation between the two controllers, thereby improving reliability while managing complexity through functional division.
3Productivity
If non-volatile RAM is used for buffering data, then write speed and data protection are improved, but memory capacity and cost increase
Solution Approach 1:
The patent changes the physical state parameter of the buffering medium by using non-volatile RAM instead of traditional volatile RAM or magnetic disk buffering. This parameter change allows the system to maintain high write speeds comparable to volatile memory while gaining the data persistence characteristics of non-volatile storage. The non-volatile nature of the RAM enables data to be retained during power failures, providing automatic data protection without requiring large capacities for redundancy, thus achieving high productivity with controlled memory capacity requirements.
Data Source
AI summary
Synchronously negotiating an object's creation time including identifying, by a first storage system among the two or more storage systems, a request to store a second updated version of the object; detecting, by the first storage system, a communications disruption among the two or more storage systems; and establishing, in dependence upon a local clock of the first storage system and a latest observed creation time associated with the object, a creation time for the second updated version of the object.