· Aug 8, 2017 1m read

Outperforming PostgreSQL and MySQL

In a previous exercise, I was able to show the power of Caché.
A medium-designed set of interdependent tables with some GB of data.
URLs cross reference over some million pages resulting in ~3 billion records

Competition was between

  • Caché
  • PostgreSQL
  • MySQL

Criteria were Speed + Storage consumption
I composed  a customized loader fed over a "raw" TCP connection
Mapping the "objects" into the final table by directly writing to Global Storage.,

Phase 1:  MySQL failed before reaching the 1st million records by it's
Incredible consumption of memory and disks space

Pase2:  Disk consumption of PostgreSQL was higher than Caché
And the load speed was remarkably lower. 
Mapping the "objects" to tables done in PHP resulting in multiple DB INSERTS 

For Caché the input was in principle a string. disassembled in the related Globals.
When Caché was done, PostgreSQL was significantly under 50% of the total.

The slim and effective design o​F Table sin Caché also showed a
dramatically better performance in data retrieval. Even with the unequal dimensions.

Finale furioso: 
The customer ordered a nicely dimensioned license for Caché.​​​​​​


Discussion (3)3
Log in or sign up to continue