Flexible Distributed MPP Partition Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional MPP database systems are inflexible, requiring the number of database partitions to match the number of fixed processors, making it cumbersome and time-consuming to add or remove partitions, as the entire setup process must be repeated.

Innovation Solution

A resource group manager initiates a management instance on an initial machine, determines the desired number of database partitions, and dynamically allocates them across available machines, allowing for flexible expansion and contraction of the resource group by adding or removing machines, and relocating management instances as needed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the number of database partitions is fixed to match the number of processors, then the system setup is simple, but the system lacks flexibility and scalability when adding or removing partitions

Engineering Contradiction:
Improveflexibility of partition managementVSAvoidcomplexity of dynamic allocation system
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic allocation by allowing the number of database partitions to be decoupled from the fixed number of processors. The system can dynamically add or remove partitions independent of processor count, enabling flexible adaptation to changing data storage needs without requiring system reconfiguration or additional hardware setup.

Inventive Principle:
Principle #15Dynamics

2Quantity of substance

If additional partitions are desired in a traditional MPP system, then data distribution capacity increases, but the entire setup process must be repeated which is time-consuming

Engineering Contradiction:
Improvenumber of database partitionsVSAvoidtime to add partitions
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-configuring the dynamic allocation framework and partition management infrastructure in advance. This allows partitions to be added or removed on-demand without repeating the entire setup process, as the underlying system architecture is already prepared to handle dynamic partition changes.

Inventive Principle:
Principle #10Preliminary action

3Quantity of substance

If the number of partitions exceeds the number of available machines, then storage capacity increases, but resource utilization becomes unbalanced

Engineering Contradiction:
Improvestorage capacityVSAvoidresource utilization balance
Core Design Contradiction:
Quantity of substanceVSEase of operation

Solution Approach 1:

The patent applies merging by allowing multiple database partitions to be combined and hosted on a single machine when the number of partitions exceeds available machines. This dynamic consolidation enables the system to maintain balanced resource utilization while providing expanded storage capacity, as partitions can be allocated across machines based on current resource availability rather than requiring a one-to-one mapping.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP2898435B1System and method for flexible distributed massively parallel processing (MPP)
Publication Date: 2018.12.12 HUAWEI TECH CO LTD
  • EP2898435B1 patent drawingFigure 1
  • EP2898435B1 patent drawingFigure 2
  • EP2898435B1 patent drawingFigure 3

AI summary

An embodiment method for massively parallel processing includes initiating a management instance on an initial machine, the management instance generating an initial partition corresponding to the initial machine, determining a total number of partitions desired for processing a database, the total number of partitions including the initial partition, determining a number of additional machines available to process the database, grouping the initial machine and the additional machines together in a pod, and launching the management instance on the additional machines in the pod to generate the total number of partitions desired for the database. Additional embodiment methods and an embodiment system operable to perform such methods are also disclosed.