With the emergence of new intelligent transportation business, various types of highway control, toll and monitoring data play an important role in various aspects of road management, and the construction of intelligent highway information technology has become the focus of highway construction. AntDB helped a provincial customer build a collaborative, shared, powerful and efficient unified intelligent management platform, which provides strong support for high-speed refined management and operation, and helps it become the first one to realize the independent and controllable upgrade of the province's high-speed clearing and settlement system.

Construction: technology builds “smart way”

　　By the end of last year, a provincial highway built mileage exceeded 10,000 km; the total number of ETC users issued in the province exceeded 6 million, the passenger ETC utilization rate exceeding 70%, and the truck ETC utilization rate exceeding 60%.

　　Since the full implementation of ETC charging on a provincial highway, human management behavior has been gradually reduced, and various kinds of tax evasion have occurred, which not only brings economic losses, but also seriously affects the management of safe passage on highway, thus the use of big data, AI and other technologies are urgently required to prevent and timely discover highway passage problems. The main challenges faced by the project are as follows:

　　· First of all, the IT information level varies from province to province, and the key information such as vehicle passage pictures on highway are kept in the center system of each road section for different lengths of time, and there are problems such as difficult data query and missing data.

　　· Secondly, the high-speed evasion of fees in provinces is mostly detected through the stage of centralized audit or relying on the experience of the chargers and auditors to conduct manual auditing, which entails many personnels with low efficiency.

　　· Furthermore, the high-speed system lacks unified management and analysis of the passing car owners, with a single dimension of analysis and poor effectiveness. The original system faces problems such as slow response at the front desk, cumbersome operation of data analysis and lack of intuitive data presentation, which makes the high-speed passing data run poorly.

Figure 1: AntDB’s business architecture of provincial road network center

　　That demands higher requirements for basic data storage architecture of highway big data platform:

　　· Unified data service, multi-mode integration: Highway big data involves many source systems, such as provincial clearing and settlement system, lane toll system, ETC gantry toll system, issuance system, customer service system, green pass system, two passenger and one dangerous cargo system, etc.; the data formats involved are diversified, such as structured passage data, passage image data, basic issued information and picture data, etc. The data storage architecture needs to be simple and easy to maintain, providing a unified access interface, storing the full amount of raw data, historical data backup, etc., and providing instant online query services.

　　· Accurate and timely response and high performance: The highway big data platform brings together up to tens of G of data collected from lane passage every day, up to nearly 100 G of basic information data obtained from the issuance system every day, and a huge amount of new customers’ vehicle permit information data and ETC gantry passage picture data, which puts forward requirements of high concurrency and low latency for data processing. At the same time, the customer requested a convenient development environment with data sharing service to complete the interoperability of file data from the Ministry's networking center and provincial centers through a nationwide unified application to guarantee the integrity and timeliness of inter-provincial and intra-provincial transaction data.

　　· Integrated management and analysis, HTAP mixed load: HTAP-type business of the provincial center system requires AntDB to meet integrated clearing and settlement and account analysis capabilities, covering business scenarios such as data collection, upload and download, data processing, clearing, clearing and settlement, data verification, path fitting, data query, etc., to achieve efficient clearing and settlement business, promote the safe and orderly development of the province's charging work on toll road network, and help refine the operation of highway.

　　Data: connect, operate and utilize

　　In the face of the system construction needs of hundreds of millions of users, traditional databases often require multiple sets of technologies to support: a set of relational databases to achieve metadata management and tagging management, and a set to manage unstructured files, with complex technical architecture, leading to difficult project management and high implementation costs. At the same time, the centralized storage architecture cannot be flexibly expanded as needed to meet the needs of business growth and massive data management.

　　AntDB has the ability of "super-converged" database, based on multi-copy data mixing technology, supporting the unified storage of multiple storage methods; at the same time, it integrates transaction processing and online analysis operations, providing one-stop data services for different data types and different demand scenarios, with low maintenance cost and high practicality.

Figure 2: the architecture of AntDB’s distributed database product

　　In the early stage of the project implementation, AntDB processed the data analysis which took 2 months before, and the load launched by the application on the database during the period was 40 times of the normal business volume, AntDB used the distributed parallel execution engine and fully utilized the hardware resources to finish the output of business reports efficiently, which was recognized by the business team and the customer.

　　During the centralized data processing, the hardware resources were basically used up:

　　· Full disk IO

　　· 80-90% of CPU usage

　　· Full 25G network, transfer rate: 2GB/s

　　After the project went online, as the business volume increased, the data volume and concurrency increased by an order of magnitude compared to the testing phase, and the market side may experience a backlog of business processing. After multiple analyses, the problem was pinpointed to disk IO performance as the key impact factor, with specific business-side performance as follows:

　　· Inbound backlog

　　· Slow recovery of wrong orders

　　· Slow gantry adjustment

　　· Slow ETC splitting

　　· Slow in splitting reports

　　During the preliminary analysis, some measures were also used for optimization:

　　· SQL optimization

　　· Table splitting

　　· Model optimization

　　There is a certain mitigation effect, but it is not enough, and there is a certain gap from the expectation.

　　After expert analysis and communication with the client, a comprehensive consideration of various aspects such as operational level, economic cost, and business impact degree, two measures were identified to fundamentally address:

　　· Host disk of database is upgraded from hard disk drive (HDD) to solid state disk (SSD)

　　· Primary and backup nodes of distributed database are deployed separately, on different hosts

　　The effect of the implemented measures is very obvious in terms of the performance of the business as follows:

　　· After the upgrade of ETC clearing in the province, the speed of business data processing has increased by nearly 15 times

　　· After the implementation of the measures, the efficiency of clearing and crediting is improved by about 60%, and the report generation time is shortened by nearly 90%

　　· After the upgrade of the network-wide ETC splitting business, the efficiency of splitting and processing of intra-provincial transactions is increased by about 3 times, and the efficiency of splitting cross-provincial transactions is significantly increased by nearly 8 times

　　· Before the upgrade, the daily backlog of gantry regulation was around one million, after the upgrade, there is no longer a backlog of historical data

　　· AntDB processes hundreds of millions of data of the clearing and settlement system every day. AntDB ensures high availability of the system by maximizing the use of hardware resources, and reduces customer costs while improving data processing performance.

Figure 3: the employment architecture of AntDB’s distributed database product

　　During a year and a half of the stable operation of the intelligent high-speed system,AntDB not only met the province’s business scenario needs for highway system of  short delivery cycle, data diversification,  multi-dimensional management, and independent and controllable technology, but also improved the the processing efficiency of real-time inbound, query and outbound of provincial highway entry/exit data. It processed the average of ten million gantry data daily, millions of lane flow, the total data volume of the database reaching nearly 100 TB, which brought into reality to "connect, operate, utilize" the massive data.

　　Application: double effect is remarkable, driving to the fast lane of intelligent transportation

　　The clearing and settlement system has been running smoothly since its launch and has achieved good economic and social benefits.

　　In terms of processing efficiency, AntDB efficiently completed the unified audit of all historical abnormal data in the past six months, and the backlog of gantry regulation before replacement was around 1 million, and there was no backlog after replacement; the data analysis work that originally took 2 months was completed in a short period of time, and the load initiated by the application during that period was 40 times of the usual normal business volume; the speed of key business processing was increased by nearly 90%, and the report generation time reduced by an order of magnitude.

Figure 4: Application effectiveness of AntDB in the industry

　　In terms of budget and cost performance, the application of AntDB domestic database has reduced millions of project construction costs and expenses, and reduced the annual commercial service costs.

　　In terms of industry demonstration effect, the successful application of domestic database for the province's clearing and settlement has set a benchmark for the application of domestic database in the highway industry and enhanced the confidence of applying domestic technology in the industry.

　　In the future, AntDB will accelerate the expansion of more digital application scenarios and continue to exert its hyper-converged database capabilities in other industry multi-business scenarios such as high-performance transaction scenarios, complex transaction scenarios, massive storage scenarios, analytical scenarios, big data analysis scenarios, etc. to provide higher-quality products and services to more industry customers.

　　About AntDB

　　AntDB was founded in 2008. On the core system of operators, AntDB provides online services for more than 1 billion users in 24 provinces across the country. With product features such as high performance, elastic expansion and high reliability, AntDB can process one million core communications transactions per second at peak, ensuring the continuous and stable operation of the system for nearly ten years, and is successfully implemented for commercial purpose in communication, finance, transportation, energy, Internet of Things and other industries.