MQSeries.net :: View topic - Dynamically controlling Aggregate Timeout in v6

bradhack · Posted: Mon Jan 22, 2007 9:17 am Post subject:

We are building a generic Enterprise Service Bus in WMB v6, and incorporating features that existed in prior middleware hubs at our company. One of them is giving the sending application the option of changing the amount of time before we send a timeout response.

We've use aggregation heavily, both in WMQI v2 and WBIMB v5 and know that the timeout parameter could only be set at design time in the Aggregate Control node. After reading some v6 documentation and help, I saw this statement from IBM:

[i]"The new aggregations nodes use WebSphere MQ message expiry to manage time outs of messages. For message expiry to work, the message queues must be browsed. The aggregation nodes browse the queues automatically to ensure that expired messages are processed."[/i]

I assume the timeout value set in the Agg Control node will set the expiry, but my question is, is it possible make a copy of the original message, overwrite the expiry based on info sent by the sending system, and send it as an Aggregate Request directly to the same reply to queue that the backend ressponses arrive on, for the fan-in process? Then, if the response messages have not arrived in x amount of time, we would format the desired timeout message appropriately.

We are not quite at v6 yet, so I cannot try a proof of concept just yet, but we need this for a detailed design document.

I know we could set up a subflow that would route to a number of different aggregate control nodes with the same control name and different timeout values, and give the users the option of selecting from a list of timeouts, but if we could allow any value, the users would prefer it.