Excellent!  

I think I would still need some serve side configuration because it
would probably be a mistake to hard code server URLs in the client
code, but that is a relatively simple task.

I am toying with sending the services-config.xml file to the client
and parsing it there.  Otherwise I am going to have to invent a
notation for defining the services.

Have you done anything along these lines?

--- In flexcoders@yahoogroups.com, "neale_patton" <[EMAIL PROTECTED]>
wrote:
>
> We do this successfully in our applications.
> 
> We have a 'shell' Flex app which loads multiple Modules -- each of
> these modules will talk to a different server-side application. Its
> not without its subleties but as Pete says creating/setting the
> channelSet manually is key to ensure consistent behaviour.
> 
> Cheers,
> Neale.
> 
> --- In flexcoders@yahoogroups.com, "Peter Farland" <pfarland@> wrote:
> >
> > Create multiple <mx:RemoteObject> instances and just programmatically
> > set a ChannelSet on each one.
> > 
> > 
> > <mx:RemoteObject id="remoteObject1" destination="destination1" />
> > 
> > ...
> > 
> > <mx:Script>
> > 
> > import mx.messaging.ChannelSet;
> > import mx.messaging.Channel;
> > import mx.messaging.channels.AMFChannel;
> > import mx.messaging.channels.SecureAMFChannel;
> > 
> > ...
> > 
> >     private function init():void
> >     {
> >         var channelSet1:ChannelSet = new ChannelSet();
> >         var channel1:Channel = new AMFChannel(null,
> > "http://server1/context/messagebroker/amf";);
> >         channelSet1.addChannel(channel1);
> >         remoteObject1.channelSet = channelSet1;
> > 
> >     }
> > 
> > ...
> > 
> > 
> > Assuming that you're only using RPC services like RemoteObject,
then you
> > don't need to compile against a services-config.xml file and can just
> > programmatically create your channels yourself.
> > 
> > This approach comes in handy for lots of scenarios - poor man's load
> > balancing by randomly assigning a server on SWF initialization, etc.
> > etc.
> > 
> > Pete
> >  
> > 
> > -----Original Message-----
> > From: flexcoders@yahoogroups.com
[mailto:[EMAIL PROTECTED] On
> > Behalf Of simonjpalmer
> > Sent: Thursday, August 23, 2007 8:59 AM
> > To: flexcoders@yahoogroups.com
> > Subject: [flexcoders] FDS mashup. Talking to multiple FDS servers in a
> > single Flex application
> > 
> > I want to use FDS/AMF to transfer data to and from my client, but
> > talking to several different servers.  I have one server for routine
> > persistence and data retrieval, a pojo server, but I have a second one
> > for generation of formatted output, a doc server.  
> > 
> > I want to report from the client state of the data, not the server
> > state, and I want to distribute my procesing so that I can scale
up the
> > architecture - doc generation can be expensive.
> > 
> > What I would like to be able to do is use the same data transport
(i.e.
> > AMF and my object model) for both the pojo server and the doc server.
> > They come together in the single Flex app.
> > 
> > Does anyone know if Flex supports this and if so how?  I seem only
to be
> > able to configure an app for a single destination.
> > 
> > OK, perhaps mashup is a bit of a stretch, but it is the same
principle.
> > I am pretty sure I could do this if I used http, but I don't want
to use
> > http, I am very happy with amf.
> > 
> > Thanks
> > Simon
> > 
> > 
> > 
> > --
> > Flexcoders Mailing List
> > FAQ: http://groups.yahoo.com/group/flexcoders/files/flexcodersFAQ.txt
> > Search Archives:
> > http://www.mail-archive.com/flexcoders%40yahoogroups.com
> > Yahoo! Groups Links
> >
>


Reply via email to