Are you sure that you are not running parallel services?
What do you mean by "mass consumption way"? Is that a single document with large number of pages? Or is it multiple OCR jobs?
"mass consumption way" --> Or is it multiple OCR jobs
multi threads call to one web Service