Data Cube High Availability via Read-Write Server Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data cube architectures lack a consistent method to ensure high availability and efficient data access, especially during hardware or software faults and heavy traffic conditions.
Innovation Solution
Implementing a separate machine for read queries and write operations, with mirrored servers for both, allowing the system to seamlessly switch between primary and mirror servers in case of failures, ensuring continuous data availability and efficient access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If read queries and write operations are handled on the same server, then device complexity is reduced, but reliability deteriorates due to single point of failure and resource contention
Solution Approach 1:
The patent divides the data cube server into separate read cube servers and processing cube servers. Read cube servers handle read queries while processing cube servers handle write operations and data processing. This segmentation eliminates the single point of failure, allowing read operations to continue even when write operations are interrupted, thereby improving reliability and data availability.
Solution Approach 2:
The patent implements mirrored read cube servers that maintain copies of the cube data. When the primary read cube server fails, the mirrored server can take over and continue serving read queries. This copying mechanism ensures high availability without requiring changes to the fundamental server architecture.
2Reliability
If read cube server is synchronized with write operations, then data consistency is improved, but productivity deteriorates due to query interruptions during synchronization
Solution Approach 1:
By separating read and write operations onto different servers, the patent eliminates the conflict between synchronization and query processing. The processing cube server performs data updates and synchronization independently while read cube servers continue to serve queries without interruption. This segmentation resolves the contradiction by allowing both data consistency and query productivity to be maintained simultaneously.
Solution Approach 2:
The patent introduces a buffer or intermediate mechanism where write operations are queued and processed by the processing cube server, which then notifies read cube servers when data is ready. This intermediary approach allows read servers to maintain their operational independence while still achieving data consistency, avoiding query interruptions during synchronization.
3Reliability
If mirrored servers are implemented for failover, then reliability is improved, but device complexity increases due to additional infrastructure
Solution Approach 1:
The patent segments the system into primary and mirrored components, where each segment has a specific role. The primary read cube server handles normal operations while the mirrored server stands by for failover. This segmentation allows the system to maintain simplicity during normal operation while providing reliability benefits when needed, reducing the perceived complexity compared to a fully redundant active-active configuration.
Data Source
AI summary
The subject disclosure is directed towards making cube data highly available and efficient to access by separating the read cube server from the processing cube server, on different physical machines. The read cube server may be mirrored, and the write cube server may be mirrored. When the primary read cube server is not operational (e.g., has failed) or is having its read cube synchronized, the read queries are handled by the mirror read cube server. When a processing cube server (or its write cube) is not operational, its mirror processing cube server takes over and performs cube processing operations via its mirror write cube.