ismaelsadeghicode/ParallelDataPartitioning

BUSINESS DATAPARTITIONING

This app implements Data Partitioning and sorting algorithm based on three disk and RAM sources and CPUs.

Technology

- Spring Boot 2.2.2.RELEASE
- Java 8
- H2 Embeded Database

Algorithm

- Datapartitioning

 - Data Partitioning
 - Redistribute
 - RoundRobin

- Sort

- ParallelBinaryMerge
- ParallelMergeAll
- ParallelPartitionedSort
- ParallelRedistributionBinaryMerge
- ParallelRedistributionMergeAll
- ParallelSort

Formol

- 1) Local merge Sort => Cost = I/O Cost + CPU Cost + Communication Cost

I/O Cost:
        I/O Cost = Load Cost = (Ri/P) * Number of passes * IO

CPU Cost = Select cost + Sorting cost + Merging cost + Generation result cost
       Select cost = |Ri| * Number of passes * (tr + tw)
       Sorting cost = |Ri| * ⌈ Log2( |Ri| ) ⌉ * ts
       Merging cost = |Ri| * ( Number of passes - 1) * tm
       Generation result cost = |Ri| * Number of passes * tw
       
Communication cost
      Communication cost = (Ri / P) * (mp + ml)

-2) Final merging => Cost = Communication Cost + I/O Cost + CPU Cost

Communication cost
      Communication cost = (R / P) * mp
      
I/O Cost:
      Save Cost = (R / P) * (Number of passes + 1) * IO
      Load Cost = (R / P) * Number of passes * IO

CPU Cost = Select cost + Merging cost + Generation result cost
      Select cost = |R| * Number of passes * (tr + tw)
      Merging cost = |R| * Number of passes * tm
      Generation result cost = |R| * Number of passes * tw