You would most likely create an Extract process with a dedicated Datapump process in the source environment. The Datapump process can suffer an outage in case of any network issue. However, it, resumes the trail file transfer to the remote site when it is up again. In such a situation you would want the Extract process to continue extracting the data into the trail files on the source system.

The trail files are stored in the filesystem which would be finite in space. The size of the trail file area will determine how many trail files can be stored, which in effect will determine how long the Extract process can continue generating the trail files while the Datapump process is down. This will determine the overall network outage capacity of your replication infrastructure. Clearly, it is quite important to size the trail file areas appropriately to avoid the situations where the Extract process has to be stopped because of lack of available space on...