| 摘要 |
A methodology for using multiple execution threads to regulate message traffic is presented. An application thread spawns a limiter thread, and each have access in shared memory to a message count that reflects a number of messages that are transmitted by the application thread within a regulated time period. The limiter thread cycles by entering a sleep state for the regulated time period, resetting the message count upon awakening, and then reentering a sleep state. The application thread executes a loop that compares the message count against a maximum number of messages for the regulated time period; if the comparison is favorable, then another message is transmitted and the message count is adjusted. If the comparison is not favorable, then the application thread enters a sleep state; after awakening, it repeats the comparison, and depending on the result, it transmits a message or reenters a sleep state.
|
