More on that "benchmark" I reported a few days ago. The guy who wrote the article I linked now has proof that MS knew about the report at least a week before its release, and had its publicity machine all wound up and ready to go. So much for impartiality.
In addition, he's found a few issues with the supposedly "optimised" J2EE database access code. Now, I'm no expert on JDBC performance tuning, but the two issues highlighted
here are just plain DUMB.
Even I could improve this stuff.And as for basing it all on a two-year-old version of Petstore...
For those of you[1] making decisions on this subject: my assessment is that the original benchmark report is effectively Microsoft-sponsored marketing
FUD. Take it with a truck-load of salt.
[1] all 0.001 of you. Or maybe I overestimate my readership. :)
[Link:
The Register]