| 摘要 |
Disclosed is a timed token protocol which may be implemented on a physical or logical ring, the ring interconnecting a plurality of stations to form a local area network. Each station on the ring may be assigned one or more of three priority levels, the priority levels corresponding to the type of service required by a station. The highest priority level (Class 1) corresponds to applications requiring a guarantee of bandwidth (e.g., digital voice), the second priority level (Class 2) corresponds to applications requiring minimum throughput, but no absolute guarantee of bandwidth (e.g., interactive data communications between terminals and hosts), and the lowest priority level (Class 3) corresponds to applications where no minimum throughput is required (e.g., batch information transfer). Class 1 stations share a pool of bandwidth, portions of which may be allocated to each Class 1 station in accordance with a bandwidth allocation procedure. A Class 1 station can transmit on every receipt of the write token the amount of information that has been guaranteed through the bandwidth allocation procedure. A Class 2 (or 3) station can transmit information upon the receipt of a write token if the time since the previous receipt of a write token is less than a first (or second) preset amount of time, the difference between the observed write token rotation time and the first (or second) preset amount of time representing the amount of unused capacity on the previous rotation of the write token. Stations assigned multiple priorities service lowest priority first, with the allowance of Class 2 priority transmission dependent on the time duration between the last receipt of a write token and the completion of Class 3 service.
|
