The mobile game company Lecastle, which has been recognized by more than 35 million users worldwide, hopes to provide better decision-making support for the game business through data analysis. Le Castle uses Amazon Redshift and other services to build its own cloud game data analysis platform on Amazon cloud technology to realize complex queries and ensure that game operators can quickly and near real-time obtain the required data and statistical information to achieve product continuity Optimization and updates to enhance the player experience. The Amazon cloud technology services used by Le Castle include Amazon Redshift, Amazon MSK, Amazon S3, Amazon CloudFront, Amazon SageMaker, etc.
Business needs and challenges
Beijing Le Castle Technology Co., Ltd. (hereinafter referred to as "Le Castle") is a company dedicated to providing high-quality mobile games for global users. The products of Beijing Lecastle Technology Co., Ltd. have been recognized by more than 35 million users around the world, and its products Alice and Wordland (Alice in Wordland) have been recommended by the Apple Store more than 3,000 times around the world.
In the operation of the game business, Le Castle adheres to the product concept of "good products must give priority to user experience, innovation should come from insight into users, details determine success or failure, and high-quality products need to be carefully polished." Excellent smartphone gaming product. To create good game works, it is necessary to better understand the preferences of players, so game data analysis is very important for Le Castle. In the past business development, different game project teams used the method of data analysis by the team itself. With the continuous development of business, enterprises need a better way to obtain global visibility.
Why Choose Amazon Cloud Technology
After evaluating different solution providers, Le Castle finally chose Amazon Cloud Technology to build its own cloud game data analysis platform based on Amazon Redshift and other services. In Lecastle's view, Amazon Cloud Technology's data warehouse and analysis solutions have the following advantages:
Purpose-built for industry needs and proven by extensive users
Amazon cloud technology provides game-specific analysis and big data application solutions for the game industry, which can help companies better understand players, develop products, improve player participation and retention rates, and allow companies to make better design decisions . These solutions are not only applied and verified by many well-known game studios, but also provide data security and compliance control for the global market, helping game companies better protect player data. "Our game business runs on Amazon Cloud Technology, including processing access logs through Amazon CloudFront, and chooses to implement data analysis of game business on Amazon Cloud Technology, which can not only obtain lower latency, but also eliminate large-scale data transmission and migration work , It also allows us to store and process various types of data in an elastic manner, reducing the construction and operating expenses in data analysis." Li Ming said.
Provide easy-to-integrate managed cloud services and real-time analysis capabilities
The importance of data analysis in the game industry to game companies like Le Castle is self-evident, so companies need to take into account the accuracy and real-time performance of data, especially for the analysis of user behavior.
Lecastle adopts Amazon Redshift products as the core data warehouse service of the enterprise. By writing embedded data of user behavior to Amazon Managed Streaming for Apache Kafka (MSK), combined with Amazon Redshift's built-in Streaming Ingestion technology, the data can be streamed in a quasi-real-time manner. The data is written to Amazon Redshift, and on this basis, subsequent required data analysis is carried out. Amazon Managed Streaming for Apache Kafka (MSK) leverages the fully managed, highly available Apache Kafka service to securely stream data. It is fully managed without worrying about infrastructure deployment, configuration, and maintenance.
It eliminates operational overhead, including the overhead of provisioning, configuring, and maintaining highly available Apache Kafka and Kafka Connect clusters. Amazon MSK makes it simple to ingest and process streaming data in real time with fully managed Apache Kafka. "This construction method simplifies the business complexity of the data analysis team, and has greatly improved real-time performance and efficiency, helping us understand user behavior faster and adjust our business strategies in a more timely manner. Data-driven business." Li Ming said.
Extensive data and analytics services to meet current and future business needs
Amazon cloud technology provides a wide range of cloud hosting services from data storage, data lake, big data analysis, streaming analysis, business intelligence and machine learning, which can meet the needs of enterprises at different development stages for data analysis and utilization. need. "One of our original intentions for establishing a unified enterprise analysis platform is to eliminate the complexity of the IT environment and business caused by each project team's self-built database and self-analysis. Combining and building, building our own data warehouse and analysis capabilities in a short period of time, so we can pay more attention to analysis strategies and rules rather than technical operation and maintenance management." Li Ming said.
Achievements
By adopting Amazon Cloud Technology's big data analysis solutions for the game industry, Le Castle not only successfully dealt with various data analysis challenges in business operations, but also achieved various results for the enterprise:
The first is to summarize the data previously scattered in various project teams, build a unified data analysis platform that serves multiple game products within the enterprise, and provide enterprises with a data-driven business operation engine. Numerous services of Amazon cloud technology provide "out-of-the-box" solutions and a user experience consistent with open source technologies. Lecastle's data analysis team can make full use of these advantages to complete the data of each project group within 2 months Integrate, migrate and build data warehouses, and design customized analysis indicators according to the needs of different businesses, freeing game project team members from complex data analysis design and operations, and transforming them into users of analysis results, so that limited Human resources focus on the innovation of game products.
Second, it allows enterprises to more confidently deal with the analytical challenges brought about by the continuous growth of the number of players. "The amount of data generated by our game business has been growing, and the business needs to observe historical data for a certain period of time. Combined with factors such as DAU, log volume, and time range, only players need to face tens of thousands of records when analyzing a scene. In the past, whether it was self-built or adopted a third-party SaaS solution, the efficiency was extremely low."
Li Ming said, "As a fully managed and highly available Apache Kafka service, Amazon MSK helps us quickly build a streaming pipeline, which improves the previous end-to-end data transmission delay from minutes to seconds, enabling real-time ingestion and processing of streaming data. Simple and efficient. Amazon Redshift, as an enterprise-level data warehouse, can accommodate data at the PB level, realize complex queries, and ensure that game operators can quickly obtain the required data and statistical information, regardless of scale or concurrent performance. We can all maintain confidence in the continued growth.”
In addition, the good integration and interoperability between various services of Amazon cloud technology also allows Le Castle to better reduce the cost of data analysis. Compared with self-built Hadoop clusters and HDFS storage, the direct connection between Amazon Redshift and Amazon S3 can simplify the architectural complexity of enterprises in the data processing and storage process and achieve better cost-effectiveness. Data engineers can use familiar ANSI SQL directly queries the open file format in Amazon S3, and can easily export various data to the Amazon S3 data lake for further processing and utilization. According to Lecastle's calculations, under the same data volume, it is more economical to use Amazon S3 to provide storage support for the data warehouse, which can help enterprises control the cost to one-tenth of the self-built method.
Looking forward to the future, Le Castle will continue to maintain close cooperation with Amazon Cloud Technology in game business operations and data analysis scenarios to better tap the hidden value of data. At present, the company has applied Amazon Redshift ML in its business, allowing developers of big data teams to use familiar SQL commands in the data warehouse to create, train and apply machine learning models in an easier way. Integrate the managed machine learning capability provided by Amazon SageMaker into the application of machine learning to complete model training and help enterprises achieve more refined business operations.