I'm using a very simple XLang schedule which (by means of
COM) extracts some data from an XML and sends this via a
WinSock channel to another machine synchronously.  When
the reply from this machine is OK, data is entered into a
database and the schedule is left.  This runs in a
transaction.

It works fine when messages are posted to BizTalk in a
once-in-a-while manner, but now, I'm putting a heavy load
onto BizTalk via a FileReceiver (could be up to 10 or more
messages dropped at the FileReceiver per seconds).

After a small while, when XLang Monitor shows that there
are 5 or more instances running, it may happen that the
oldest messages are not threaded anymore.

Does anyone have a clue how to overcome this ?

Thanks. John.

[ Post a follow-up to this message ]

I think the problem you are trying to solve is how to "throttle" the
orchestration instances so that they do not overwhelm the processing
system.

If this is the issue, then you might want to attach the orchestration
to a XLANG schedule pool which can be controlled (via instantiation
counters). Do not forget to appropriately set the wait time too.

XLANG schedule pool is available within COM+ services.
To attach the particular orchestration to XLANG schedule pool, open
the BEGIN icon and attach the orchestration to the particular pool you
want.
Hope this helps.

"John" <anonymous@discussions.microsoft.com> wrote in message news:<061d01c3dab2$b2dba9c
0$a001280a@phx.gbl>...
> I'm using a very simple XLang schedule which (by means of
> COM) extracts some data from an XML and sends this via a
> WinSock channel to another machine synchronously.  When
> the reply from this machine is OK, data is entered into a
> database and the schedule is left.  This runs in a
> transaction.
>
> It works fine when messages are posted to BizTalk in a
> once-in-a-while manner, but now, I'm putting a heavy load
> onto BizTalk via a FileReceiver (could be up to 10 or more
> messages dropped at the FileReceiver per seconds).
>
> After a small while, when XLang Monitor shows that there
> are 5 or more instances running, it may happen that the
> oldest messages are not threaded anymore.
>
> Does anyone have a clue how to overcome this ?
>
> Thanks. John.

[ Post a follow-up to this message ]