ws-soap-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bill Wishon <>
Subject RE: Transaction ID Namespace (was Object lifetime options)
Date Thu, 02 Nov 2000 20:26:41 GMT
The only real requirements I have are:

One client be able to manipulate many server side objects without those
objects having to be serialize back and forth across the network.

If no one accesses a particular object for a certain amount of time that
object is discarded from the server side object pool.

The word transaction was just a word I picked up while reading the soap
specification.  It seemed to me like what the soap specification was
describing as a transaction ID was similar to what I wanted to accomplish.
In my original e-mail about implementing a new object lifetime model I had
called it an object ID.  With a transaction/object ID in the headers of the
soap message the body of the soap message will still look like a regular
soap RPC message.  After the first RPC message to the server the client
remembers the transaction ID that comes back in the response header.  From
that point on whenever the client uses that ID in the header it knows which
object on the server side it is going to manipulate.

While looking through the Apache soap code I found the available scope
options of application, session, page and request insufficient for what I
want to accomplish so I embarked on creating a transaction scope.  Will this
new pluggable provider code satisfy the requirements I have?  Instance IDs
seem like just another name for what I was calling a transaction/object ID.
View raw message