Pumpkin Film is a membership-only video platform focused on offering personalized, high-quality movies and TV dramas to its 80 million users, differentiating itself from traditional online video platforms. Through an ad-free model with no hidden or additional costs, it provides users worldwide with high-quality movies and TV shows by creating a pure, user-friendly viewing environment.

The deep integration of Big Data, virtual reality and other emerging technologies with video media has enabled the Internet entertainment industry to undergo rapid digital transformation in recent years. Data shows that the number of online Chinese video users is increasing year on year, with the number of users reaching about 734 million in 2021.

The steady increase in the number of online movie viewers has led to many theatrical releases achieving record growth with movie and TV platforms becoming the main vehicles for meeting the increasing demand for movie experiences.

The prosperity of the industry also means that the platform must attract viewers by offering more innovative services than traditional offline theaters. Pumpkin Film is committed to creating an immersive, bespoke viewing experience for moviegoers using data. Each movie on the Pumpkin platform is recategorized via a new intelligent algorithm based on data analysis and each user’s unique profile, creating an accurate list of recommended movies personally tailored to each user, so that the content of the platform and the user's viewing habits intelligently feed into each other. In addition, a new Cinema Room feature allows certain users to set up their own Cloud Cinema Room and invite friends and family to gather virtually for an interactive movie watching experience.

New virtual movie theaters are not able to control or predict the number of viewers in advance in the same way that brick-and-mortar movie theater operators can, so they often face unpredictable peaks and troughs in traffic. In order to ensure a smooth viewing experience for all users, the database must be flexible and resilient. Also, to meet increasing demand for real-time interactions and chat, Pumpkin Film added new features and games, including pop-ups comments from viewers, which means it’s essential the database can accommodate fast iteration to support ongoing future progress.

Pumpkin Film relied on a traditional MySQL relational database in the early days, but as the number of users grew, the required maximum number of individual tables began to put a strain on the database. Pumpkin Film found that when the data in a single table exceeded five million, the performance of the database decreased greatly, and when the data in a single table reached 10 million, the database could only perform a small part of any simple query function. When it came to adding or changing a new attribute given increasing users, the traditional database was not fit for the task, taking over an hour to add any new attribute. This caused extreme inefficiencies in development and iteration.

Given the constraints of the relational database implementation, Pumpkin Film first tried to split the table and split the library, but this solution increased performance risks for the database. The coupling with code created more pressure as the code had to handle the logic of splitting the library and splitting the table to ensure normal business operation. This jeopardized both the code and the database which caused serious problems for the business.

After being evaluated by the team, Pumpkin Film ultimately chose AsparaDB for MongoDB. MongoDB easily handles splitting libraries and tables. It also supports distribution which improved database flexibility and scalability, thus increasing development efficiency while reducing the complexity of daily operations and maintenance.

AsparaDB for MongoDB is a fully managed cloud database service operated by AliCloud and certified by MongoDB Inc – the company that develops and maintains the MongoDB software. In 2022, MongoDB and AliCloud celebrated their three-year collaboration anniversary. In November 2022, the latest version – AsparaDB for MongoDB 6.0 – was launched to more effectively enable Pumpkin Film to optimize modern application architectures, support a wider variety of scenarios, and drive business performance and scaling advancements.

Take the "log system" – the largest application scenario of AsparaDB for MongoDB at Pumpkin Film – as an example. The log system introduces protocols from various channels, including business logs and application logs resulting in a flexible and malleable data structure. This is something the previous traditional database would struggle with. As the business develops, the logs grow rapidly, affecting data storage, and requiring a greater need for flexible expansion according to demand. At the same time, log storage must be smoothly integrated into the Big Data framework.

Facing highly competitive data scenarios with up to millions of writes per second, AsparaDB for MongoDB now provides stability and optimum performance. First, the scalability and high availability of the system have been greatly improved. Based on MongoDB's distributed architecture, Pumpkin Film can now more easily handle the horizontal expansion of Big Data scenarios with no down time ensuring a smooth user experience. Second, the flexibility of MongoDB supports rapid business changes while improving developer efficiency. Developers can now introduce more new game features much faster which also enhances user experience. Adding attributes to a single table with 10 million data points in a traditional database previously took over an hour, but it now occurs instantly with MongoDB.. The release schedule has been accelerated from the previously slow pace of once a month to once every two weeks ensuring new applications are launched on time.

Finally,day-to-day operations and maintenance now function with greater ease for Pumpkin Film’s 200 employees.With rapidly increasing users base and business growth, AsparaDB for MongoDB has simplified system monitoring, backup and recovery, freeing up developers' time, and making dealing with unexpected situations faster and more efficient.

“What impressed us most with AsparaDB for MongoDB is its flexible and variable data structure and its ability to handle Big Data scenarios,” said Yin Jinliang, head of big data and infrastructure security, Pumpkin Film.

“The drawbacks of traditional relational databases are at odds with developers' 'open principles.' MongoDB is very developer friendly. It has freed us up to focus on actual business innovation without spending too much time on the database structure, and without worrying about code issues affecting the database performance and causing risks. Our experience reinforces that AsparaDB for MongoDB is significantly better than a traditional relational database in terms of development efficiency and performance."

Yin Jinliang, head of big data and infrastructure security, Pumpkin Film