After the "Tianqing" business operation,
What are the differences between the familiar meteorological business
?
builds a unified, open and intensive platform, bringing massive data and complex business systems of the meteorological department together, allowing information to be unimpeded, and the integration and cooperation between observation, forecasting and services has become unprecedentedly convenient and efficient. Many years ago, these ideas were just "castles in the air", but now, with the business operation of the meteorological big data cloud platform, it is reflected in reality one by one. "Tianqing", a "digital and computational integrated" support platform, is becoming the most reliable pillar of the "cloud + end" new business format with highly intensive meteorological business and R&D.
Growth
Fight the difficulties and build a solid "cloud + end" business technology base
record the changes in the cloud with observation data, and then obtain the forecast through complex operations, and finally transform it into a rich and diverse service product. The meteorological business chain is highly dependent on the smooth flow of data and information from beginning to end. Establishing an intensive, stable and efficient information system has always been the ideal of meteorological workers.
entered a new era. With the rapid development of information technology and the increasing desire of meteorological business to have intensive and efficient systems, "Tianqing" was born because of the situation and shouldered the important task of consolidating the technical foundation of the "cloud + end" business.
01
was born because of the situation
December 2017 is the starting point of the construction of "Tianqing".
On December 11, the "Meteorological Big Data Action Plan (2017-2020)" was released, proposing to uniformly design and build an open and interconnected meteorological big data cloud platform. On December 26, the "Meteorological Information Development Plan (2018-2022)" was issued, clarifying the construction of a meteorological big data cloud platform with the "proprietary cloud + public cloud" model.
Hi Wenchun, deputy director of the System Engineering Office of the National Meteorological Information Center, participated in the entire process of "Tianqing". He believes that from the initial stage, the positioning of "Tianqing" has been very clear, and "intensive" is the most obvious feature. , a new platform, will gather, share, open and apply meteorological data resources, promote meteorological applications into it, build a smart meteorological ecological system, support the open sharing of meteorological data resources and cross-field integrated application, and improve the level of intensification and application collaboration capabilities of meteorological information systems.
At that time, the National Comprehensive Meteorological Information Sharing Platform (CIMISS) had just solved the problem of applying unified data sources in the country and province, but the "information island" and "application chimney" problems between business units and business systems have not been resolved, and resource waste is prominent. How to achieve higher-level intensiveness on the basis of data intensiveness? The design of "Tianqing" needs to be carefully considered.
From the national level to the provincial level, the meteorological business system is very complex. To smoothly integrate different systems, we must be targeted and guided by needs. In January 2018, the National Meteorological Information Center, in conjunction with national-level business units and some provincial-level meteorological bureaus, formed a 34-person design team to begin designing a meteorological big data cloud platform. After the first draft of the design was completed, feedback was collected comprehensively and 82 suggestions for improvement were absorbed. It took 9 months to complete the "Meteorological Big Data Cloud Platform Design Plan". The construction of "Tianqing" has been officially launched.
02
overcome difficulties
0 For decades, meteorological departments at all levels have built hundreds of different business systems such as data specifications, system architectures, programming language , etc. according to their respective needs. It is not easy to build a compatible platform.
Unlike the Internet "big manufacturers" to build cloud platforms, the construction of meteorological big data cloud platform faces severe algorithm compatibility problems . Taking numerical forecasting as an example, my country's model research and development began in the 1980s, and many core algorithms still use Fortran language, which was born in 1957 or C language issued in 1972.To reconstruct these algorithms from the bottom layer, it is neither realistic nor necessary, so it is necessary to enhance compatibility and allow programs from the "old era" to run smoothly. Compared with the compatibility of individual algorithms, the overall integration of business systems is a more systematic project. He Wenchun introduced that the biggest reason why the phenomenon of "standing chimneys" in meteorological business systems was widely present in the past was that each system had different needs and could not be fully met by a single model. As the first business system integrated into the operation of "Tianqing", the Climate Monitoring, Predictive Analysis System (CIPAS) has such a unique need.
for climate monitoring and prediction analysis requires a lot of meteorological data marked with geographic information. According to the five storage technologies designed by Tianqing at that time, processing this data requires the extraction of meteorological data and geographical information and then superimposing the two. For CIPAS, such a model will lead to reduced business operation efficiency. Therefore, the technical team communicated fully with the business personnel, adding a spatial analysis library for "Tianqing" to store its own geographical information and meteorological data. In this way, CIPAS's business needs are met, and "Tianqing"'s support capabilities have been further improved.
was born on the Northwest artificial weather operation command system on "Tianqing" in the "cloud native" posture, and faces completely different challenges. Since subsequent new business systems will adopt the "cloud native" model, this system developed directly on the cloud needs to be "produced" for later generations. After unified development, the authority of the "standardization" of the Northwest artificial weather-affected operation command system was fully demonstrated, and it took less than three weeks to complete the standardized deployment in Inner Mongolia, Gansu, Shaanxi, Qinghai, Ningxia and Xinjiang.
demand and R&D promote each other. With this model, "Tianqing" released the national-level version 1.0 in July 2019, with the capabilities of massive data storage, full service integration, and efficient data application. Next, this platform needs to be spread across the country.
03
Joint growth
0Deploy "Tianqing" in provincial meteorological departments is not an easy task. On the one hand, "Tianqing" is built on a brand new big data and cloud computing technology, with a high technical threshold; on the other hand, "Tianqing" is a huge scale as the base of the new business technology system. Since the completion of CIMISS, provincial technical strength has been strengthened, but there are still shortcomings in the deployment of "Tianqing".
Faced with this situation, the technical team adopted an open and cooperative approach to vigorously introduce provincial technical personnel to develop and build, and 50 technical personnel from provincial meteorological departments participated. They came with demand and technology, becoming the backbone of the smooth deployment of "Tianqing" at the provincial level. In December 2019, 31 provincial meteorological departments completed the deployment.
Close cooperation between the country and the province, and national-level business units were also fully linked, and 7 people participated in the construction with demand. Within the National Meteorological Information Center, the system engineering room responsible for development and construction is more intimate in cooperation with the data service room responsible for user services and the operation and maintenance of the operation and monitoring room.
As the "Tianqing" is taking shape step by step, talents are also growing. Since entering the construction of "Tianqing", 2 people in the team have been hired as second-level professional positions, 3 people have been promoted to the senior professional title ; 5 people have been selected as high-level talents of in in , including 1 meteorological leader, 1 chief meteorological expert, 2 young meteorological talents, and 1 outstanding meteorological talents in the western and northeast China; 4 people have been hired as deputy commander-in-chief or deputy chief engineer in key projects of the China Meteorological Administration, 10 people have been hired as chief engineer or deputy chief engineer in the first-level system; 8 people have been selected as high-level talents of the National Meteorological Information Center or provincial bureau.
On December 15 this year, "Tianqing" officially started its business after nearly a year of trial operation, but the construction of "Tianqing" in is far from over. most national and provincial business systems need to be integrated, and the application of advanced technologies such as artificial intelligence on the big data cloud platform needs to be solved. In the future, the technical team cultivated during the construction process will become the most important driving force for "Tianqing" to further grow.
advantages
"Integrated Numbering" What changes will bring?
At 8:00 on December 15, the meteorological big data cloud platform "Tianqing" was officially operating in business, providing "digital and computing integrated" platform services to the whole country. Because "Tianqing" performed well in the trial operation, this switch seemed "silent" because it was reasonable, but what hidden under the calm surface is the all-round and full-chain upgrade brought by "Tianqing" to the meteorological business.
01
promotes the organic integration of observation and forecasting services
takes forecasting accuracy as the leader, driving the development of precision monitoring and fine service. To achieve such a development model, "Tianqing" is an indispensable foundation.
Various business systems are integrated into "Tianqing", which not only means that it operates here and obtains data. The data and products generated during business operations can also be standardized through "Tianqing". In this way, the distance between various observation, forecasting and service business systems is unprecedentedly narrowed . The forecasting business system can obtain various observation data more efficiently, the service system can better utilize forecast conclusions, and the sharing of data and collaborative cooperation between various systems can be seamlessly connected.
02
greatly enriches data resources
"Tianqing" provides complete data resources. As of December 15, "Tianqing" has fully integrated CIMISS's original data, historical archived data, various data sets and submission data, significantly increasing the length of domestic and foreign data time series; 611 data data, with 221.49 billion records, with 296 million new daily additions; the global ground, high altitude and ocean century sequences were first extended to 1901, the global aircraft report sequence was extended from 4 years to 43 years, and the radar-based data sequence was extended from 2 years to 13 years. Compared with CIMISS, the data integrity and data quality of the meteorological big data cloud platform have been improved to varying degrees, and the data metadata information is more complete and of higher quality.
"Tianqing" provides efficient data services that integrate real-time history. Application system builders or developers do not need to find data from various sources, establish collection channels, data cleaning, and parse storage. They just need to apply for a data usage account and select the data required by the application to access efficiently.
03
Improve data utilization efficiency
"Tianqing" Through upgrading and perfect data storage technology, the time cost of business operation has been further reduced. is based on message middleware, data flow, file transmission, satellite broadcasting and network crawling technology. "Tianqing" has established a multi-protocol-supported data exchange system, which can actively discover and automatically collect various types of data, and can be processed while collecting it.
In view of the characteristics of meteorological business, "Tianqing" has built a distributed storage system to reorganize data for applications. For data with high real-time requirements such as weather forecasting, a distributed table system is adopted to have access time of milliseconds; for climate equal-length sequence analysis applications, a distributed analytical database is adopted to have analysis time of seconds. In addition, for self-use data such as intermediate results and experimental data, personalized storage is supported, and storage space is allocated at one time and used freely.
04
Reduce application development costs
"Tianqing" provides rich algorithms and computing resources, which can effectively reduce application development costs and shorten the online cycle. As an application system builder or developer, there is no need to purchase, install, deploy and maintain hardware resources. It only needs to submit an application integration plan and apply for computing resources that the algorithm depends on. At the same time, there is no need to develop all the algorithms required by the application. By applying to use the algorithm shared in the algorithm library, algorithm multiplexing and sharing by the public intelligence are realized.
For the storage resources required for product processing in the application, by applying for the "Tianqing" application special database, there is no need to build an independent database on its own, which greatly reduces operation and maintenance costs and improves data sharing levels, which is conducive to joint analysis of multi-source data.The processing and generation of products in the business not only requires corresponding data and computing power, but also requires the establishment of a task scheduling framework. The processing pipeline provided by "Tianqing" realizes a variety of task engine scheduling, including time triggering, data source triggering, event triggering, etc. The algorithm can configure tasks by selecting the corresponding scheduling engine. The application only needs to pay attention to the algorithm implementation logic.
integrates
comprehensively "clouded" efficiency several times
Since the construction of "Tianqing", business systems of meteorological departments at all levels have successively carried out integration work, and efficient synergy benefits have been initially revealed. At present, "Tianqing" operates 7×24 hours a day, with the reliability of continuous operation at the national and 31 provincial levels exceeding 99.9%, the success rate of task scheduling at the national and provincial levels reaches 100%, and the scheduling delay is controlled at the millisecond level.
Climate Monitoring Prediction and Analysis System (CIPAS)
Integration: The system architecture is fully "clouded" and the system performance is comprehensively improved. After the integration of
CIPAS, the system architecture is fully "clouded", realizing the comprehensive integration of data, processes, platforms and monitoring. The system is centered on the forecaster experience, and its performance in data access, processing, algorithm operation, graphics visualization and product generation has been comprehensively improved. The average data query time has been reduced by more than 50%, and more than 70% of the products that require long-term calculation and timed generation have been changed to real-time interactive acquisition; CIPAS 3.0 national first-level deployment and business trial.
Northwest region artificial weather-affected operation command software system
integration situation: "cloud native" development, standardized deployment is convenient and efficient.
This system has completed the research and development of the "cloud native" model based on "Tianqing" and completed unified and standardized deployment in 6 provinces. During the period of the 14th National Games in , based on the stable computing and storage resource support provided by "Tianqing", the system successfully completed the guarantee task.
Intensive grid forecasting new process
integration: plays a key role in real-time correction of forecasts and forecast consistency.
Based on the "Tianqing" national and provincial "one network", National Meteorological Center and Beijing and Hebei have built a new intelligent grid forecasting process from grid forecasting to data application. Based on this new process, the 2022 Winter Olympics competition area and surrounding cities site forecasts are processed and produced based on this new process. From the correction and release of forecasters to the application system to obtain revised forecast information, the entire process takes only tens of seconds, playing a key supporting role in real-time forecast correction and forecast consistency.
Comprehensive meteorological observation business operation information platform ("Tianyuan")
integration situation: data and algorithms are intensively managed, and redundancy is reduced by 30%.
"Tianyuan" is the first core business system to complete the transformation and integrate into "Tianqing". After integration, the intensive capacity of data and service resources has been increased by more than 50%, and the product processing timeliness has been increased by more than 30%, supporting the efficient operation of the observation business and providing reference demonstrations for subsequent project integration; realizing unified service of metadata "Tianqing".
area high-resolution numerical forecast inspection and evaluation system
integration: data reading efficiency is effectively improved, promoting the intensive management of algorithm .
This system is completely based on the development and construction of "Tianqing", realizing traditional inspection, spatial inspection and process evaluation of regional numerical forecast precipitation, as well as near-ground element inspection. The data reading efficiency is effectively improved, realizing direct storage and direct access of mode data and observed data. Based on the "Tianqing" product processing assembly line, we will effectively promote intensive algorithm management and complete real-time monitoring of system software and hardware.
radar estimation precipitation system
integration: system process optimization, timeliness has been nearly doubled.
realizes radar-based data reading from "Tianqing", algorithms based on processing assembly line scheduling, radar product writing back "Tianqing", and full-process monitoring based on "Tianjing".After the system is integrated into "Tianqing", the process is better, reducing the data transfer process, direct data source access and product storage, and the timeliness of product and service are nearly doubled.
Ningxia Intelligent Comprehensive Meteorological Service Sharing Management Platform
Integration: Data storage timeliness has doubled, and business system query and statistical data efficiency has been increased by 3 to 4 times.
data, algorithms, systems, monitoring, etc. are simultaneously integrated, and the core business system realizes the reconstruction of functional modules. Relying on the storage and computing resources of "Tianqing", the operation efficiency of the business system has been significantly improved, the timeliness of data entry into the database have been doubled, and the efficiency of business system query and statistical data has been increased by 3 to 4 times.
Hubei Province Meteorological Forecast Service Business Integration Platform
Integration: More than 100 business service products are connected, and 82 algorithms are divested into the "cloud".
core system is fully integrated into "Tianqing", eliminating self-built databases, accessing more than 100 business service products, and divesting 82 algorithms into "cloud". After integration, the business process has become more standardized, and data timeliness, algorithm operation efficiency and hardware resource utilization have been significantly improved.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-