Solution highlights: analysis on hundreds of billions of spatio-temporal data records of vessel trajectories, integration of multi-dimensional spatio-temporal data, and shortened development cycle.
After we began to use Ganos offered by Alibaba Cloud, our capabilities to process heterogeneous spatio-temporal data have greatly improved. Thanks to the high performance, low latency, high throughput, and high stability of ApsaraDB for Lindorm (Lindorm), we have saved lots of O&M costs and can focus more on implementing solutions. ––Suggested by the Chief Technology Officer (CTO) of Elane Inc.
Elane Inc. is the largest data service provider in terms of the automatic identification system (AIS) in China. The company provides leading big data services for the shipping industry in China. The company aims to speed up the transformation and upgrade of the shipping industry by using big data technologies to drive the integration of global shipping logistics and the Internet. The company is providing services for millions of global users from diverse industries, such as companies from the port, shipping, logistics, finance, insurance, and device manufacturing industries. The core features of a visual analysis system for shipping data are to analyze the shipping data that customers are concerned with and provide analysis reports. In most cases, heat maps and statistical metrics are included in the analysis reports. For example, Lindorm can be used in the following scenarios:
Analysis on the crowdedness of routes that are used to transport bulk commodities
Distribution of routes for the transport of iron ore, liquefied natural gas, or other cargo
Analysis on the crowdedness of ports
Research on vessel traffic volumes of multiple routes based on the cross section of a waterway
Research on the frequency of periodic activities by vessels in a region
Research on how a policy affects the business in a region
High-efficiency data access is required. If the performance of data queries is low, the user experience of the visual data analysis system is significantly compromised.
The data volume is large and data extraction is time-consuming. The AIS data from Elane Inc. has been accumulating for more than 17 years, with a daily increment of about 1 billion data records. Elane Inc. needs to analyze the historical data that was generated over a long period of time.
The technologies used to implement business solutions are complex. The business of the company involves many types of data which is stored by using different storage engines.
The Ganos engine of Data Lake Analytics (DLA) encapsulates the code to read and write spatio-temporal data in a unified manner to simplify code compilation.
The Ganos engine is also used to efficiently process raw data to reduce the time spent on data extraction and data analysis.
The extracted data is stored in Lindorm Ganos. Lindorm features low latency, high performance, and high throughput. Its excellent query capabilities meet the business requirement for smooth visual display of business data.
The following figure shows the architecture of the visual analysis system for shipping data.
The following figure shows the system architecture.
Alibaba Cloud services help Elane Inc. shorten the cycle of implementing technologies in business and reduce O&M costs.
Ganos on ApsaraDB for Lindorm excels in the metrics of data query performance. This helps you meet the business requirements for low latency and visual display of shipping data.
Shipping route data is different from trajectory data in that shipping route data is more relevant to business. Therefore, additional methods must be used to store and query shipping route data. To respond to this need, Ganos of DLA allows the company to integrate various types of spatio-temporal data and analyze the data.
Data in the shipping industry is complex and custom computing operations on the data are required. To meet this requirement, Ganos of DLA provides a built-in Spark cluster. In this cluster, an engine is available for customized data processing. This way, the company does not need to separately purchase a Spark cluster.
Ganos of DLA supports pyramid storage that enables a quick display of vector data. This optimizes the display of resizable heat maps.