I am amazed that folks who consider themselves computer professionals
would say things like "always" when talking about performance. For
example, I'd expect that a JVM in hardware (or with a hardware assist)
would run faster than a conventional machine running C. Even trying to
cover yourself by saying "well, I was assuming conventional hardware"
doesn't make it -- what hardware, what software? What application? What
does the XML look like? Would base64 decoding help (e.g., most
xml-dsig's)? There's just too much variation out there.
Performance comparisons are very hard to do correctly and must be properly
qualified.
I assert that folks making blanket assertions should reconsider their
claims. :)
/r$
--
STSM
Senior Security Architect
DataPower SOA Appliances
