Redeliver undelivered messages that are NOT covered by the Retry Policy mechanism

Experience operating a B2B Advanced Communications/ IBM MQ / Sterling Business Integrator ebms AS4 gateway has shown that there are many ways in which outbound messages can fail. Relatively few of these are covered by the Retry policy mechanism built into B2BAC.

There examples not covered by Retry are -

1. HTTP 500 connection error

2. Socket time out error

3. Connection forcibly closed by remote host

The attachment shows just show of 40,000 messages over the last 12 months that have required expensive manual user intervention and reprocessing to simply send the message again. In most cases these were simply the receiving partner being busy or temporarily unavailable or an unstable network connection. Our 5 years of experience operating this software. If we extrapolate the cost out over the last 5 years - the inefficiency and cost and reputational impact is significant.

From an operational perspective there is no difference between a message that was not able to delivered due to a B2BAC destination being marked as unavailable and in retry mode as opposed to one of the 3 failure types indicated above. A message failure is a message failure regardless of the mechanism by which it failed.

This request comes of the back of a PMR which recently advised us that the 3 scenarios identified above are NOT covered by the Retry Policy functionality. Consequently, we require a feature built in B2BAC where it recognises outbound failure (including the cases above) and will attempt to redeliver those messages again automatically. This would reduce considerable manual overhead and expense associated with supporting messaging through this product.