axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sanjiva Weerawarana" <>
Subject Re: [Axis2] - OM and MTOM
Date Thu, 28 Oct 2004 17:56:39 GMT
"Aleksander Slominski" <> writes:
> >Alek, the MTOMness is only known via the WSDL.
> >
> interesting - so MTOM is bound now to WSDL?

MTOM is not bound to WSDL. However, the way you declare that
optimization is supported by a service is by indicating that in the
service's WSDL.

> if SOAP payload is in multipart-mime HTTP message would it not be strong
> enough (sufficient?) indication that it may be  MTOM?

Yes of course. However that's not the case we're talking about. Take
an element foo:
Is that an integer, a string or a base64 encoded binary byte array? That
cannot be determined without schema (WSDL) info.

I believe that's the case that Eran was addressing. However, if the XSD
Media Types annotation has been applied then one may be able
to tell whether its the last case or not:
    <foo xmlmime:type="image/jpeg">1234567890</foo>

The point Eran was raising was that looking for this attr always is a perf
cost. I personally think that's a valid perf cost .. like looking for
which also we have to do. (I may have thought differently earlier but
now I'm convinced ;-).)

> moreover on server side when service is deployed we can indicate if it
> supports MTOM or not - right?

Yes whether to MTOMize or not.

> so in this case WSDL is only needed on client side to automatically turn
> on MTOM support?

MTOM is an *optional* *optimization*. The WSDL must tell the client whether
the service accepts this optimization or not (and whether it will generate
it or not).

> still i think on client side we can have in SOAP2 API way to manually
> turn MTOM on?

Sure. However we do want that API properly configured from the info
in the WSDL too, right?

I think we're agreeing ...


View raw message