Distributed Hashing Storage with Supervisor Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chord-based network technologies lack effective fault tolerance and content management in cases of node joins, leaves, or failures, and do not provide a way to automatically rebalance or repair the network in the event of temporary node absences or content misplacement.
Innovation Solution
A distributed hashing mechanism with transactional support and failover capability is implemented, using clear assignation systems for node IDs, chunk IDs, and replica IDs, along with supervising computers for automatic rebalancing and detecting chunk range overlaps to ensure data availability and integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If Chord-based network is used for distributed storage, then scalability and decentralized routing are improved, but fault tolerance and automatic rebalancing capability deteriorate
Solution Approach 1:
The patent implements a feedback mechanism where the supervisor system continuously monitors node status and triggers automatic rebalancing operations. When nodes join or leave the network, the supervisor detects these changes and initiates content redistribution to maintain even load balancing across the network, thereby providing fault tolerance while preserving scalability
Solution Approach 2:
The patent introduces a supervisor system as an intermediary that coordinates node operations and manages content placement. This supervisor acts as a mediator between nodes and the distributed storage system, handling failover scenarios and rebalancing operations without requiring nodes to have complex fault tolerance logic, thus maintaining scalability while adding reliability
2Quantity of substance
If content is distributed across multiple nodes, then storage capacity and availability are improved, but data integrity and consistent access deteriorate
Solution Approach 1:
The patent implements preliminary actions by pre-calculating content placement locations using consistent hashing algorithms and pre-establishing replica configurations. When data is stored, the system proactively determines which nodes should hold which content and in what quantities, ensuring data integrity is maintained from the outset rather than requiring complex verification later
Solution Approach 2:
The patent uses parameter changes in the form of content identifiers and node identifiers passed through hashing functions to dynamically determine data placement. By changing the input parameters (content ID, node ID) through consistent hashing, the system ensures that data integrity is maintained across distributed nodes while allowing flexible storage capacity expansion
3Device complexity
If manual node management is used, then system simplicity is improved, but operational efficiency and automatic recovery deteriorate
Solution Approach 1:
The patent implements self-service mechanisms where nodes automatically register themselves with the supervisor system upon joining the network. The supervisor then automatically assigns content and configures replicas without requiring manual intervention. This maintains system simplicity from the user perspective while dramatically improving operational efficiency through automated node management and self-healing capabilities
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A multipurpose storage system based upon a distributed hashing mechanism with transactional support and failover capability is disclosed. According to one embodiment, a system comprises a client system in communication with a network, a secondary storage system in communication with the network, and a supervisor system in communication with the network. The supervisor system assigns a unique identifier to a first node system and places the first node system in communication with the network in a location computed by using hashing. The client system stores a data object on the first node system.