Modular Pipelined Analytics for Multitenant Data Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multitenant databases face inefficiencies in processing complex data analyses due to the computational load of commingled data, as the commingled data server struggles to handle the combined data from multiple tenants effectively.
Innovation Solution
A system that determines and executes computational steps in subsets, using both tenant-associated processors and commingling-associated processors, adhering to pipelining criteria, to efficiently process and analyze data by breaking down computations into preprocessing, main processing, and postprocessing subsets, and transferring data appropriately between these processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all computational steps are executed by the commingled data server, then data analysis can be performed centrally, but the computational load becomes overwhelming and processing efficiency deteriorates
Solution Approach 1:
The patent divides computational steps into three distinct subsets: preprocessing steps executed by tenant-associated processors, main processing steps executed by the commingled data server, and postprocessing steps executed by tenant-associated processors. This segmentation distributes the computational load across multiple processing units, preventing any single server from becoming overwhelmed while maintaining coordinated data analysis.
2Adaptability or versatility
If tenant data is commingled for analysis, then valuable cross-tenant insights can be generated, but data security and isolation are compromised
Solution Approach 1:
The patent implements different processing locations for different computational stages: tenant-associated processors handle preprocessing and postprocessing of sensitive tenant data, while the commingled data server handles main processing with commingled data. This ensures that tenant data remains under tenant control during critical stages while still enabling cross-tenant analysis in the main processing stage.
3Productivity
If computational steps are distributed across multiple processors, then processing efficiency improves, but system complexity and coordination overhead increase
Solution Approach 1:
The patent performs preliminary actions by executing preprocessing steps before main processing, where tenant-associated processors prepare and filter data before it is commingled. This preliminary preparation reduces the complexity of subsequent main processing by ensuring data is in the appropriate format and only relevant data is commingled, simplifying the coordination required in later stages.
4Reliability
If the commingled data server processes all data, then data consistency is maintained, but processing time and resource consumption increase significantly
Solution Approach 1:
The patent enables continuous useful action by allowing tenant-associated processors to independently execute preprocessing and postprocessing steps without waiting for the commingled data server. Meanwhile, the commingled data server processes main processing steps in parallel. This continuous parallel execution maintains data consistency through coordinated interfaces while significantly reducing total processing time.
Data Source
AI summary
A system for performing a computation includes an interface and a processor. The interface is configured to receive an indication of a computation from a tenant associated user. The processor is configured to determine a set of computational steps for performing the computation, where each computational step of the set of computational steps adheres to a pipelining criterion. The processor is configured to determine associated processors for subsets of the set of computational steps, where the associated processors include a tenant associated processor or a commingling associated processor. The processor is configured to cause execution of the set of computational steps using the associated processors.