-------
The XML is a compound object.
The XML is store in SQL Server, and send by email to other city.
An app regenerate the object from the XML and store it in SQL Server in relational mode.
---------

Are you saying that the XML is a persisted object description? For example, a setup object for an application, where each file in the setup is represented as a binary blob directly in the XML?

That's what it sounds like you're doing.

If it were me, and if it were anything like this, I would forgo the binary blobs in the XML.

I do understand how you're e-mailing the thing but the XML could contain a UNC/URL reference to the blobs so they could be d/l'd as part of the setup process rather than contained directly in the XML. Or your setup could contain references to a web service that would make the files available on demand during the setup. Your setup definition would be a rather small XML file <s>.

I'm only guessing, and the Setup scenario described here is only an example <s>.

But, think of it this way: If you have a very large object, sometimes it is not efficient (like a monolithic app), just like in relational databases, and you have to look to the design. Even with security requirements, and even though you have to make this stuff available to a remote location, this is not the only way.

>L<