China's leading game company Lilith Games (Lilith Games) and the volcano engine ByteHouse have reached a cooperation to work together to accelerate the real-time data warehouse construction of Lilith Games. This cooperation will use ByteHouse's innovative technology and functions to provide comprehensive support and assistance for Lilith's DAP efficiency improvement.
Lilith Games is a representative of Chinese Mesozoic game companies and maintains a leading position in the Chinese game market. In order to support its growing business needs, Lilith Games has widely used ClickHouse for advertising placement and operation analysis in China, and built an open source ClickHouse cluster in overseas business to support advertising operation analysis. However, some challenges have been encountered during these usages, including cross-month and cross-season query timeouts, OOM, lack of support for large table associations, and poor stability of real-time synchronization from MySQL data. Therefore, Lilith Games hopes to solve these pain points by investigating the product functions and performance of ByteHouse and exploring cooperation opportunities.
As the core cloud data warehouse product of Volcano Engine, ByteHouse has many powerful functions and provides comprehensive support for the real-time data warehouse construction of Lilith games. First of all, ByteHouse's MaterializedMySQL function can synchronize the entire database of business data in Lilith game MySQL or PolarDB MySQL to the ByteHouse platform in real time. This function provides a visual management tool and adds a convenient exception operation and maintenance processing function to achieve the goal of one-click synchronization and stable operation. This greatly reduces the cost of data synchronization and the workload of operation and maintenance, and provides an efficient data synchronization link for Lilith games. In particular, ByteHouse's unique HaUniqueMergeTree table engine supports real-time table-level/partition-level unique key update, partial column update, real-time deletion, etc., allowing applications to easily solve real-time data deduplication and update problems.
Secondly, ByteHouse's self-developed query optimizer solves the problem that the open source ClickHouse does not support complex query capabilities. The optimizer supports RBO, CBO, and distributed plan optimization, and implements high-level optimization capabilities such as Dynamic Filter pushdown, materialized view rewriting, and cost-based CTE. Through these capability optimizations, ByteHouse was able to successfully run all 99 SQL use cases of the TPC-DS test set, and greatly improved performance. In the actual environment, ByteHouse helped Lilith game advertising business to improve the performance of complex query analysis scenarios by more than 3 times, and supported the query time span from 1 month to more than 1 year, helping it cover more analysis business scenarios and speed up decision-making and optimization business operations.
In addition, ByteHouse also provides visual operation and maintenance functions. Aiming at the high complexity of ClickHouse operation and maintenance, it provides a series of visual functions such as cluster management, monitoring and alarming, health evaluation, authority management, parameter configuration, data management, and interface expansion. . This greatly reduces the operation and maintenance cost and provides a more convenient and efficient operation and maintenance experience for Lilith games.
Through the cooperation with ByteHouse, Lilith Games will build the core OLAP engine of the advertising operation analysis platform based on this platform. This will enable Lilith Games to process massive delivery data and log data in real time, and perform high-performance computing. Through real-time statistics and attribution analysis of the full link data of advertising delivery, Lilith Games can adjust the advertising delivery strategy in real time to improve the effect of buying volume. At the same time, through fine-grained cost and effect analysis, Lilith Games can support R&D, operations, publishing and other teams to carry out refined operation and management, better understand market demand, and enhance collaboration efficiency and transparency. In addition, Lilith Game also uses ByteHouse to support the function of crowd packs, which can quickly screen out users who meet specific conditions in massive data scenarios, and achieve precise targeted advertising.
This cooperation will provide Lilith Games with strong data capability support and help it maintain a competitive advantage in the game industry. The two parties will work together to promote the development of real-time data warehouse construction and bring more efficient and innovative solutions to business fields such as advertising operation analysis.
About Lilith Games
Founded in 2013 and headquartered in Shanghai, China, Lilith Games is a well-known game company integrating game development and global distribution. Lilith Games is committed to creating high-quality game content and constantly promoting the innovation and development of the game industry. At present, Lilith Games has many best-selling game products around the world, and is actively expanding overseas markets.
About ByteHouse
ByteHouse is a cloud-native data warehouse under Volcano Engine, focusing on providing real-time data warehouse solutions for enterprises. ByteHouse helps enterprises build an efficient, stable and scalable real-time data analysis platform through innovative technologies and functions, provides powerful data processing and analysis capabilities, and helps enterprises stand out in the fiercely competitive market.
Click to jump to the volcano engine ByteHouse official website to learn more