Four construction of Yangshan Deepwater Port processing power required to break through the traditional system of terminal business to achieve the transformation from the centralized to a distributed architecture, and the pier operating system for smart scheduling software service (ITF) is the core software component pier production management system , undertook the task of receiving the entire production system, intelligent task scheduling assignment, power station operator intelligent scheduling, return results after the task is completed, the cluster management and operational audit.
Due to the sheer size of Yangshan Deepwater Port, automation equipment range, ports scheduling assignments increasing, and the presence of the peak scheduling, resulting in large-scale data processing and data storage capacity, high speed and accurate data acquisition (5 ms / time) , thus stabilizing the intelligence service task scheduling software for reliable, self-healing fault tolerance, response time, throughput becomes critical, requires intelligence service task scheduling software (ITF), distributed transaction database software platform service components do online rolling upgrade.
Is more than a good deal with the problem, we can understand the solution heat Pu database HotDB offer:
hot Pu database team conducted a distributed architecture design and transformation, technology architecture core set of systems used in all domestic software and open source software such as: development language JAVA, operating system is CentOS, application server Tomcat and Nginx, big data platform for Hadoop, database HotDB distributed transaction database, server hardware is made in China.
2, designed to achieve the overall structure and function of ITF intelligent scheduling system to solve the unmanned intelligent scheduling production management system task group of the whole terminal. ITF intelligent scheduling system uses a service oriented architecture, ITF at least three-node cluster, horizontally scalable. Online and be able to service the ITF Server upgrade, upgrade selected node in the cluster node list, separated from the node to the cluster, automatic upgrade, upgrade success of the node is greater than half of clustering upgrade the bulk of (the non-upgraded nodes for the bulk separation, the upgraded node joins the cluster batch), then upgraded and non-upgraded node added to the cluster.
3, the establishment of two data centers, data center and disaster recovery primary data center. The main data center to deploy high-availability version of the ITF Management management platform, the cluster version of the ITF Server service, load balancing cluster, Redis cluster. The primary data center services for daily business. Deploy full-service backup data center single-point version, as disaster recovery.
4, according to the Yangshan Deepwater Port task scheduling service software (ITF) technical requirements for distributed transaction database, database thermal Pu were carried out technical architecture design architecture design data persistence layer, data caching layer, the overall deployment architecture design, multiple stages of research and development work function and high-performance protection, intelligent management platform R & D functions and high-performance protection and other back-end applications, the successful implementation of the requirements of the system architecture, high availability, high throughput, high-performance, stable operation and so on. Business system to meet the automated terminal annual throughput of 4,000,000 containers, transport equipment 150 an hour 400,000 assignments of demand, guarantee the stable operation of the entire business.
Task scheduling management platform
combined with the customer's application, the actual value of the thermal Pu database solutions summarized from several aspects:
1, industrial-grade high availability data services and application services guarantee high availability of business systems and stable, reliable and security; continuously provide services to the business system enables operators to dock more smooth and efficient.
2, intelligent scheduling pier operating system task group, autonomous intelligent application and database failure to achieve the unmanned automatic operation and maintenance management business system architecture, help architecture Yangshan Port and unmanned technology walking in the same industry the most front end.
3, from centralized to distributed transformation used to achieve a distributed transaction database HotDB persistent mass data storage, online capacity expansion and the level of intelligent operation and maintenance of the database level, effectively reducing the threshold for the use and operation and maintenance costs.
4, data caching layer Sentinel architecture reduces the risk of system downtime, improve business continuity systems and services, thus providing a higher level of stability data caching services.
5, in response to the call of domestic self-control, the use of core application components are completely self-development, open-source in-memory database.
Line time on the project: 2017 - Present
Data Node Scale: 64 data nodes
The maximum number of concurrent operations: 152 000
Traffic data: 96TB / 6 months
Pier automation equipment: 760 and above
Data acquisition period: 1 / Device / 5 ms
