To calculate the total disk space necessary to handle all queues during a worst-case failure (4 hours), use the formula:
Sum(InboundQueueSizes) + Sum(OutboundQueueSizes)
where:
InboundQueueSizes are the two queues calculated in “Inbound queue size example calculation”.
OutboundQueueSizes are the three queues calculated above.
The total disk space needed in the worst case, including 2MB for the RSSD inbound queue, is:
2MB + 28MB + 8MB + 151MB + 324MB + 194MB = 707MB
Sybase recommends that you allocate enough space for your worst case scenario. If you use the save interval feature (outbound queues are not truncated even after messages are delivered to the next site), you must be sure to allocate enough queue space to sustain your peak transaction activity. If you do not use save interval, then under normal circumstances, your queue utilization will be very small, perhaps 1MB or 2MB per queue.
The example calculations assumed that all outbound queues would have to tolerate the same duration of failure. This assumption may not be true in your environment. Typically, connections across a WAN must tolerate longer duration failures than local connections.