Has anyone done any performance comparison’s with large Odata data sets (ie Parttran, GLJrnDtl)?
Versus a direct SQL query on a local machine I’m seeing Odata as 5-1 to 10-1 slower on data retrieval from a BAQ.
Anyone else?
Has anyone done any performance comparison’s with large Odata data sets (ie Parttran, GLJrnDtl)?
Versus a direct SQL query on a local machine I’m seeing Odata as 5-1 to 10-1 slower on data retrieval from a BAQ.
Anyone else?
BAQ is always going to be slower than direct SQL because of the many layers in between. BAQ adds security checks, plus all the WCF Services noise.
Add to that REST which is yet another layer on top of it and you get
SQL -> WCF -> REST Facade -> Network -> You
So yeah it will be slower than direct SQL, a more “fair” comparison would be a BAQ via BO/WCF vs a BAQ View REST.
In my experience the difference between those two methods usually throws the advantage to the REST because it is much more paired down than the BL directly.
I guess that explains why EDA gets to put a local application on the Epicor hosted servers to extract the data via SQL, instead of using the oData they force the ‘masses’ to use to access their SaaS data.