DNS Query Prediction
Lu, Dai (2024)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202404055749
https://urn.fi/URN:NBN:fi:amk-202404055749
Tiivistelmä
DNS query event as the prerequisite of network connection, introduces extra time latency to complex network application which involves a group of network activities.
The latency cannot be optimized in the pipeline of network interactions in current DNS framework. This thesis proposed a prediction solution and corresponding learning method to reduce the overall latency. By estimating conditional probability, which is a widely used metric in natural language processing to solve the “word association” problem, which is similar to this problem, this thesis proposed a learning method to learn query associations from DNS traffic. This thesis also proposed a standalone prediction solution and an integrated prediction solution to cooperate with current DNS cache mechanism, to accelerate DNS service by predicting.
The latency cannot be optimized in the pipeline of network interactions in current DNS framework. This thesis proposed a prediction solution and corresponding learning method to reduce the overall latency. By estimating conditional probability, which is a widely used metric in natural language processing to solve the “word association” problem, which is similar to this problem, this thesis proposed a learning method to learn query associations from DNS traffic. This thesis also proposed a standalone prediction solution and an integrated prediction solution to cooperate with current DNS cache mechanism, to accelerate DNS service by predicting.