Background
Map World is a network geographic information sharing and service portal constructed by National Geomatics Center of China (NGCC). It integrates a wide range of data sources to provide authoritative, standard geographic data through national e-government networks.
With about 600 million daily visits to Map World, the current system was unable to keep up with increasing service requirements. There was insufficient O&M manpower, so the O&M was difficult. To reduce costs, they used cloud services provided by public cloud vendors and tailored to their increasing service requirements.
Challenges
As services grew, the MongoDB Community Edition was unable to quickly respond to the levels of concurrency Map World services needed to handle. They needed a database that could process 600+ million concurrent requests per day.
As services grew, the MongoDB Community Edition was unable to quickly respond to the levels of concurrency Map World services needed to handle. They needed a database that could process 600+ million concurrent requests per day.
As the number of tile layers and data volume increased, it was not easy to scale out MongoDB. A new database was required to handle at least 20 TB of tile data.
As the number of tile layers and data volume increased, it was not easy to scale out MongoDB. A new database was required to handle at least 20 TB of tile data.
The O&M workload was increasing. Three persons were required for maintaining devices, two for services, and one for security. Automated database O&M was needed to ensure stable performance.
The O&M workload was increasing. Three persons were required for maintaining devices, two for services, and one for security. Automated database O&M was needed to ensure stable performance.
Solutions
GaussDB(for Mongo) helps promote geographic information sharing and cloud-based innovation of Intelligent data services
The data import bottleneck was eliminated and standby nodes no longer hang when tiles are loading.
The solution uses the hybrid cloud architecture, which combines Huawei Cloud for external services and Map World private cloud for internal tests. First, Huawei Cloud GaussDB(for Mongo) was used to process tile data for online maps. Compute nodes can now be added in minutes and storage capacity can be scaled out in seconds to process high-concurrency requests. Then RDS for PostgreSQL was used to process vector and 3D data, ensuring data reliability and integrity. Finally, RDS for MySQL was deployed to manage users and theme layer attributes. Databases are fully managed on the cloud, allowing customers to focus on application development without worrying about database O&M.
GaussDB(for Mongo) supports 500+ million daily visits, ensuring stable operations with zero downtime.
? Low costs: GaussDB(for Mongo) uses snapshots to implement read/write splitting, reducing the cost by half.
? High availability: Stateless shard servers support failover in seconds, ensuring high availability.
? Easy O&M: Automatic O&M is enabled based on management and control capabilities.
? High performance: GaussDB(for Mongo) is optimized based on RocksDB and provides higher performance than on-premises MongoDB databases. GaussDB(for Mongo) meets all customer service requirements.
Benefits
Huawei Cloud databases provide reliable and efficient services for Map World
