| 摘要 |
The invention concerns a method, a computer program product and a network entity (ePDN GW / HA, Fig. 4 ) for handling fragment packets, which fragment packets (FP, Figs. 3a , 3b , 8 ) respectively comprise parts of a data packet of a data stream ("first flow" or "second flow" in Fig. 4 ), which data packet comprises data packet header information ("UDP header", Fig. 3a , 3b ) and which data packet is longer than a maximum allowed data packet length defined by a protocol used to transmit the fragment packets,
wherein the network entity comprises:
-a buffering unit (BU) for temporarily buffering fragment packets arriving at the network entity,
-a reassembling unit (RU)
--for reassembling data packet header information ("UDP header", Fig. 3a ) from one fragment packet (FP#1. Fig. 3a ),
if the complete data packet header information is in this one fragment packet (FP#1),
--for reassembling the data packet header information ("UDP header", Fig. 3b ) from a plurality of fragment packets (FP#1, FP#2, Fig. 3b ),
if the data packet header information is distributed over a plurality of fragment packets
(FP#1, FP#2, Fig. 3b ),
-an interface to a fragment packet filtering unit (EPDN / GW or GW (HR) or GW (FR) in Figs 4 , 8 ) for filtering fragment packets depending on reassembled data packet header information.
|
