Greg Hess <[EMAIL PROTECTED]> writes: > For fun I tested RMI too: > > Case: 100M zip, over 1000 calls with zip byte[] passed as method arg: > > Product Memory(ave per > call) Time(ave round trip) > > Java RMI > 100M(480Mpeak) 4033.62/ms > > > > Axis XML parsing consumes less memory that RMI object serialization and > has almost the same round trip time.
I'm interested and confused. Were you not testing a 5MB zip with SOAP serialization? Why use a 100MB zip with RMI? I'm generally very curious about how SOAP in general, and Axis in particular, stacks up against RMI and Java serialization, so if you ever do get around to doing some more extensive testing, I'd be interested in hearing about it. As would plenty of other folks as well, I'm sure. -- joe
