| 摘要 |
<p>There exists a certain class of packet messaging problems where a number of competing yet equally important packet message sources are all transmitting the same type of packet message and these packet messages are to be sent over such a low bandwidth link that congestion will occur. The result of this congestion will be that not all the packet messages sent from the number of packet message sources will be able to be carried on the link. One example of such a situation is a distributed game architecture where entities hosted in a virtual environment on, for example, a server computer, transmit packet messages containing updates as to their respective positions in the virtual environment. These packet messages might then be transmitted over a low bandwidth to the client computer of a game player. If a present day scheme, such as a layered video coding type scheme, were applied to select which packet messages could be carried on the link, then the client would have a good deal of knowledge about the state of some entities but very little knowledge about the state of others, which may be a critical disadvantage. According to the invention, packet messages are sent with an associated packet message priority setting on a scale of 0 to n, in sequences of n + 1 packet messages cycling through each priority setting in the priority setting scale. In this way, if packet messages are dropped from the sequence due to congestion on a priority basis, the remaining packet messages of the sequence as left equally spaced as is possible. Thus, as the link bandwidth capacity is exceeded and packet messages begin to be dropped, or rather queued, on a priority basis, the packet messaging method and apparatus according to the invention ensures that, as far as possible, regular update packet messages are still received from all packet message source entities.</p> |
