Distributed Data Storage with Symmetric Peer Nodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems face limitations in accessing and retrieving large amounts of data due to hardware reliance, bandwidth constraints, and inefficiencies in data transfer, leading to bottlenecks and limited processing capabilities.
Innovation Solution
A distributed data storage system with interconnected peer node computer systems, using symmetric storage applications and locally attached storage, where data objects are identified by unique object identifiers and stored in a flexible manner based on reliability criteria, allowing for distributed recovery and enabling applications to be executed directly on the storage system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is stored on disk drives with rapidly increasing storage capacity, then storage capacity is improved, but the ability to locate, access, and retrieve selected data is limited by disk head performance and communication bandwidth
Solution Approach 1:
The patent segments data into discrete objects with unique identifiers, allowing parallel retrieval of multiple objects simultaneously across distributed storage nodes. This eliminates the sequential access limitation of traditional disk heads and enables concurrent data fetching operations.
Solution Approach 2:
The patent transitions from centralized sequential data access to distributed parallel access across multiple storage nodes. By distributing data objects across different locations in the storage system, retrieval operations can proceed simultaneously across multiple dimensions (nodes), bypassing the single disk head bottleneck.
2Reliability
If highly configured data storage systems with costly hardware are used to manage large data storage systems, then storage reliability is improved, but device complexity and cost increase
Solution Approach 1:
The patent implements data replication where copies of data objects are stored across multiple nodes. This distribution of copies provides redundancy and reliability without requiring complex hardware configurations, as the reliability mechanism is achieved through simple copying and distribution of data.
Solution Approach 2:
The system automatically manages data distribution, replication, and retrieval without requiring complex hardware configurations or specialized administrators. The distributed architecture enables self-service operations where nodes autonomously handle data storage and retrieval tasks.
3Stability of the object's composition
If lock managers are used to prevent concurrent modification of stored data, then data coherence is improved, but the lock managers become a bottleneck in the data storage system
Solution Approach 1:
The patent segments data into independently addressable objects with unique identifiers, allowing concurrent access to different objects simultaneously. This eliminates the need for global locks and enables parallel data operations without interference, as each object can be modified independently without affecting other objects.
Solution Approach 2:
Instead of using locks to prevent concurrent access and then managing complexity through lock management mechanisms, the patent inverts the approach by allowing concurrent access through content-addressable identification. The system achieves data coherence through the inherent properties of content addressing rather than through restrictive locking mechanisms.
4Ease of operation
If data is transferred from the data storage system to the user's system for application execution, then user processing capability is improved, but data transfer time increases due to bandwidth limitations
Solution Approach 1:
The patent merges the data storage system with the application execution environment by enabling applications to run directly on the storage system nodes. This eliminates the separation between data storage and processing, allowing computations to be performed where data resides without requiring time-consuming transfers to remote systems.
Solution Approach 2:
The patent introduces content-addressable unique identifiers as intermediaries that enable direct access to data objects stored on the system. These identifiers allow applications to locate and access specific data objects efficiently without requiring full data transfers, serving as a mediator between data storage and processing operations.
Data Source
AI summary
In accordance with the invention, a data storage system for fixed content includes one or more peer node computer systems interconnected by a storage system network, where each peer node computer system includes a symmetric storage system application and locally attached storage. A data object is distributively stored on one or more of the one or more peer node computer systems and identified using a unique object identifier. A data object is flexibly stored on the data storage system according to external criteria. Stored data objects are accessible from any of the one or more peer node computer systems. Applications can be executed on the data storage system using data objects that are stored on, being written to, or being read from the data storage system to generate results accessible by a user.