Data Grid Replication Using Versioned Single-Node Backup
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data replication in data grids leads to increased latency due to the need for multiple communications between nodes, which can be unsuitable for certain users, processes, or applications, and requires two predesignated backup nodes for storing data.
Innovation Solution
Implementing a method where an originator node transmits a write command to a single predesignated backup node, which stores the data and returns a success message with a version number, allowing the originator node to store a local copy, thereby reducing latency to half that of traditional methods and using only one backup node.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is replicated to multiple predesignated backup nodes, then fault tolerance is improved, but latency increases due to multiple communications
Solution Approach 1:
The patent extracts the version number information from the data replication process itself and uses it as a separate mechanism to track and manage replicas. By separating the version tracking function from the data transmission function, the system can identify and contact only the most current replica without requiring communication with all predesignated backup nodes, thus reducing latency while maintaining fault tolerance
Solution Approach 2:
Instead of having backup nodes proactively store and manage multiple copies of data, the patent inverts the approach by having the originator node track version numbers and selectively replicate data only when necessary. This inversion transforms the system from one where backup nodes continuously hold data copies to one where replication occurs on-demand based on version number comparisons, reducing unnecessary communications and latency
2Reliability
If two predesignated backup nodes are used for data replication, then fault tolerance is improved, but device complexity increases
Solution Approach 1:
The patent makes the originator node universal by enabling it to perform both data generation and replica tracking functions. The originator node maintains version numbers for all replicas and determines when replication is needed, eliminating the need for separate predesignated backup nodes with specialized roles. This multi-functionality reduces device complexity while maintaining fault tolerance through dynamic replica management
Data Source
AI summary
Data replication in a data grid can be managed. For example, a computing device usable in the data grid can determine a version number for a key-value pair. The computing device can determine the version number by increasing a counter corresponding to the key-value pair to a new value and using the new value as the version number for the key-value pair. The computing device can store the key-value pair and the version number for the key-value pair in a memory device that is included in the computing device. The computing device can then transmit an electronic communication to a remote computing device that is located remotely from the computing device in the data grid. The electronic communication can cause the remote computing device to store the key-value pair and the version number for the key-value pair in another memory device that is included in the remote computing device.