Kazuya Kimita

Decreasing FCD Processing Delay by Deploying Distributed Processing System

A field test for floating car data (FCD) based traffic information processing & provisioning system is carrying on in Beijing, with over 2,000 floating cars uploading FCD every 90 seconds. However, as a practical system targeted by the government, 20,000 or more floating cars will be deployed for FCD collecting. Ten times more amount of FCD might lead to the unacceptable processing delay. Multiple servers based distributed FCD processing system is a cost efficient way to achieve the low processing delay and high availability. This paper suggests the structure of distributed system and algorithms of optimizing the distributed system by efficiently balancing work load, which can effectively keep the processing delay within an acceptable level