Multi-source Logistics Data Management Architecture

As the logistics platform gained in popularity, extreme volume of logistics data has been generated and is continuously growing in size. It becomes a pain point to efficiently query and analyze for different sources of logistics. However, most of the existing distributed methods aim at managing a single-source data like spatial data or trajectory data and build spatial-temporal indexes to improve query efficiency. Thus it is in urgent need of designing an efficient data management architecture for supporting query or analysis on multi-source logistics data. On the basis of distributed environment, we first split massive logistics data into partitions in terms of time dimension, and apply hash algorithm and broadcast mechanism for each partition to accelerate data fusion. Further, we obtain multi-attribute trajectories by regarding the property of other sources of data as the attributes related to the trajectories, and build a distributed index to proliferate the efficiency of querying for logistics data. Finally, comparative experiments are conducted to demonstrate the advantages of our proposal, and a demo system is built for a logistics platform to showcase the effectiveness of our proposal.