Hi Thomas, it probably isn't the SAAJ implementation alone but as well the XML parsing components that cause more problem than they solve. Right now we've isolated those key items: 1. Chunked transfers 2. Encapsulation namespaces with single quotes Neither of them are against XML/HTTP rules, but they casue a large number of users of our services to fail due to interoperability issues (a key importance to us). 1. Chunked transfers We were able to remove this by adding a couple of additional classes to handle IO (which just shouldn't be needed). 2. Namesapces Using those specialiced classes we were able to resolve this issue as well. Overall a lot of our subscribers are running embedded systems that rely on minimal overhead for parsing XML and thus will sometimes choose a parser that is not 100% compliant but allows them to save memory footprint. Our objective is to make sure they work. SAAJ (and the required XML parser) were far more tolerant on receiving as well as providing a wider interoperability than what we now have. Regards, Reiner View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3988676#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...