performance - What XSLT processor is better to use for small parallel transformations -

performance - What XSLT processor is better to use for small parallel transformations -

i need execute lots (abot 100-200) of parallel xml transformation in cycle. average xml size 5kb. relation of used xsl xml parsed 0.25, means xmls sharing same xsl. xsl here cached, unfortunately xsl not stored in file, , generated dynamically within application. xslt processor suits improve case?

p.s. language - java, transformation final result type - string

thanks in advance

i'm not sure have understood question, i'll try.

firstly, there seem 2 separate questions, think orthogonal. first question of how cache stylesheets: if running transformation more 1 time it's thought compile 1 time , utilize repeatedly. in general isn't problem, unless there isn't plenty memory cache stylesheets in case need kind of lru strategy.

the sec question selection of xslt processor. think can utilize same caching architecture whichever processor choose, question won't constrain choice. in java world main free/open-source processors xalan (with versions apache , embedded in jdk) , saxon-he; since saxon-he supports xslt 2.0 , faster there little competition. among paid-for processors main contenders ibm's websphere processor , saxon-ee; should both handle workload ease, you're unlikely consider ibm product unless you're making substantial investment in ibm middleware.

(i won't seek hide i'm developer of saxon product...)

performance xslt