Fault Tolerant Reservation State Distribution for Multi-Partition Volumes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing block-based storage systems face challenges in maintaining a consistent reservation state across multiple partitions hosted by different data storage units and attached to multiple clients, leading to inefficiencies in distributing reservation state changes without requiring additional connections.
Innovation Solution
A fault-tolerant mechanism that uses existing connections between clients and servers to distribute a shared state, where a designated server stores an authoritative copy of the reservation state, and clients poll and broadcast the current state to other servers and storage units, ensuring consistent views without additional overhead or latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If storage reservations are made between a single server and attached storage, then the reservation mechanism is simple to implement, but it does not function properly for block-based storage volumes comprising multiple partitions hosted by different host devices and attached to multiple clients
Solution Approach 1:
The patent divides the storage volume into multiple partitions hosted by different host devices, with each partition maintaining a copy of the reservation state. This segmentation allows the reservation mechanism to scale across multiple servers while maintaining consistency through coordinated state management among partitioned storage units.
2Reliability
If additional connections are established to distribute reservation state changes across multiple partitions, then consistency can be maintained, but operational costs and system complexity increase
Solution Approach 1:
The patent makes existing storage connections serve dual purposes: both data transfer and reservation state distribution. By utilizing the same connection infrastructure for multiple functions, the system maintains consistency across partitions without requiring additional dedicated communication channels, thereby reducing operational complexity.
Solution Approach 2:
The system enables partitioned storage units to automatically distribute and synchronize reservation state changes among themselves using existing connections. The storage system self-manages the state distribution without requiring external coordination infrastructure, reducing both complexity and operational overhead.
3Loss of energy
If existing connections are utilized for distributing reservation state, then operational costs are reduced, but the mechanism must efficiently maintain consistency across multiple partitions without additional latency
Solution Approach 1:
The patent implements continuous state synchronization by having partitioned storage units maintain and propagate reservation state changes continuously through existing connections. This continuous action ensures that state information flows seamlessly across the distributed system without interruption or additional latency, as the same connections used for data operations also handle state distribution.
Data Source
AI summary
A storage service supports attachment of multiple clients to a distributed storage object and further supports persistent reservations that govern types of access the respective clients are granted with respect to the distributed storage object. In order to efficiently distribute reservation state changes to multiple partitions of the distributed storage object hosted by different data storage units/servers, existing connections are used between the data storage units/servers hosting the partitions of the distributed storage object and the connected clients to propagate reservation state changes.