DLI is applicable to large-scale log analysis, federated analysis of heterogeneous data sources, and big data ETL processing.
Large-scale Log Analysis
- Gaming operations data analysis
Different departments of a game company analyze daily new logs via the game data analysis platform to obtain required metrics and make decision based on the obtained metric data. For example, the operation department obtains required metric data, such as new players, active players, retention rate, churn rate, and payment rate, to learn the current game status and determine follow-up actions. The placement department obtains the channel sources of new players and active players to determine the platforms for placement in the next cycle.
- Advantages
- Efficient Spark programming model: DLI directly ingests data from DIS and performs preprocessing such as data cleaning. You only need to edit the processing logic, without paying attention to the multi-thread model.
- Ease of use: You can use standard SQL statements to compile metric analysis logic without paying attention to the complex distributed computing platform.
Federated Analysis of Heterogeneous Data Sources
- Digital service transformation for car companies
In the face of new competition pressures and changes in travel services, car companies build the IoV cloud platform and IVI OS to streamline Internet applications and vehicle use scenarios, completing digital service transformation for car companies. This delivers better travel experience for vehicle owners, increases the competitiveness of car companies, and promotes sales growth. For example, DLI can be used to collect and analyze daily vehicle metric data (such as batteries, engines, tire pressure, and airbags), and give maintenance suggestions to vehicle owners in time.
- Advantages
- No need for migration in multi-source data analysis: RDS stores the basic information about vehicles and vehicle owners, table store saves real-time vehicle location and health status, and GaussDB(DWS) stores periodic metric statistics. DLI allows federated analysis on data from multiple sources without data migration.
- Tiered data storage: Car companies need to retain all historical data to support auditing and other services that require infrequent data access. Warm and cold data is stored in OBS and frequently accessed data is stored in GaussDB(DWS), reducing the overall storage cost.
- Rapid and agile alarm triggering: There are no special requirements for the CPU, memory, hard disk space, and bandwidth.
Big Data ETL Processing
- Carrier big data analysis
Carriers typically require petabytes, or even exabytes of data storage, for both structured (base station details) and unstructured (messages and communications) data. They need to be able to access the data with extremely low data latency. It is a major challenge to extract value from this data efficiently. DLI provides multi-mode engines such as batch processing and stream processing to break down data silos and perform unified data analysis.
- Advantages
- Big data ETL: You can enjoy TB to EB-level data governance capabilities to quickly perform ETL processing on massive carrier data. Distributed datasets are provided for batch processing.
- High Throughput, Low Latency: DLI uses the Dataflow model of Apache Flink, a real-time computing framework. High-performance computing resources are provided to consume data from your created Kafka, DMS Kafka, and MRS Kafka clusters. A single CU processes 1,000 to 20,000 messages per second.