We find out the forwarding an entire study to EFILM ist mutch more stable the forward picture by picture immedetly. The problem with "Wait xxx minutes for all instances of the study to be received" is that if the minutes value is to high the doctors complain that the get the examination to late, if the value is too low, pasone sends the hole study again and again, witch cause high load and hangs on the destination AE.
The idea is not to forward the entire study after a certain time, but the entire series immedetly after that moment when pacsone recieves an new series number for this patient.
E.g. patient study #1 starts to recieves pictures from CT with series id 101. Pacsone does not forward anything. When the CT starts to send series 102 pacsone recognize that the series 101 is finished and send the entire series to the destination AE. The same happend when Series 103 starts to recieve with series 102.
With that you would have the best of both worlds. the doctors get the series complet and as soon as possible.
autoroute: forward entire series immediately
That's a great idea, but still there are 2 potential problems:
1. It assumes the sending Dicom application will always send all instances of a study in order, i.e., one series followed by another. But the current Dicom standards have no requirement for such ordering, e.g., the sending application can send Instance#1 of Series#1, followed by Instance#2 of Series#2, followed by Instance#2 of Series#1, etc, and still comply fully with the current Dicom standards. Such out-of-order transmittance of the instances of a Dicom study would break the implementation of this feature.
2. Even if we can assume all instances of a study are transmitted by series order, i.e., Series#1 followed by Series #2, etc, we still face the same problem as the Wait N minutes and forward the entire study option, as we still need an indication of when all instances of the series/study have been transferred, before we can schedule a routing job to forward the entire series/study. In other words, we can start forwarding the previous series as soon as the next series starts to come in, but we still don't know when the last instance of the last series has been received. This is the same problem as we don't know when the last instance of the entire study is received, thus the need for the Wait N minutes and forward the entire study option.
1. It assumes the sending Dicom application will always send all instances of a study in order, i.e., one series followed by another. But the current Dicom standards have no requirement for such ordering, e.g., the sending application can send Instance#1 of Series#1, followed by Instance#2 of Series#2, followed by Instance#2 of Series#1, etc, and still comply fully with the current Dicom standards. Such out-of-order transmittance of the instances of a Dicom study would break the implementation of this feature.
2. Even if we can assume all instances of a study are transmitted by series order, i.e., Series#1 followed by Series #2, etc, we still face the same problem as the Wait N minutes and forward the entire study option, as we still need an indication of when all instances of the series/study have been transferred, before we can schedule a routing job to forward the entire series/study. In other words, we can start forwarding the previous series as soon as the next series starts to come in, but we still don't know when the last instance of the last series has been received. This is the same problem as we don't know when the last instance of the entire study is received, thus the need for the Wait N minutes and forward the entire study option.