How Important Is the Database in Game Development?
How Important Is the Database in Game Development?
Data-driven insights drive user experience, which drives revenue. In this interview, learn how important databases are in game development.
Join the DZone community and get the full member experience.Join For Free
MariaDB TX, proven in production and driven by the community, is a complete database solution for any and every enterprise — a modern database for modern applications.
Thanks to Ben Ballard, Customer Success Manager at VoltDB for sharing his insights on the current and future state of game development. VoltDB provides an in-memory, translytical database that’s able to ingest millions of transactions per second. They have several large game development companies as clients.
Q: What are the keys to developing successful games?
A: A successful game has always needed engaging gameplay, but to retain and monetize players over the long term, the most successful games leverage data to constantly improve the game and personalize the experience of each player. This means not only capturing game play data but also tracking players in real-time, and the ability to bring data-driven insights into the real-time personalization of the game. That’s why it’s important to build a data analytics platform that runs concurrently with the game.
Q: What are the most significant changes to game development?
A: The sophistication of the games and their ability to personalize the UX. Crossy Road, developed by Hipster Whale, is a great example of a game that adapted the gameplay of Frogger from the 80s with new characters and a variety of endlessly scrolling worlds to become successful in 2014. It has had impressive staying power by introducing coins and tickets that are earned as you play, when you return to the game, and can also be purchased. Spin-offs such as Crossy World Disney provide additional characters and worlds to unlock. To better understand the players, Crossy Road games uses a platform provided by Game Analytics, which is powered by VoltDB. This helps them segment players by a number of different behaviors and attributes, as well as purchase history, provides this summary to the game in real time and enables the game to make better offers for improved player retention and higher conversion rates on in-game purchases.
King Entertainment, makers of Candy Crush and its many sequels, along with many other games, has a large data infrastructure on a global scale capturing metrics in real-time from millions of people playing concurrently. Using Kafka to stream data into VoltDB, King is able to keep everything they know about a player accessible and current so that it can be leveraged by the games. They are perform deep analytics on this data over time to cluster players into groups to maximize retention and gameplay. When the game loads, they can retrieve the history of the player, which segment(s) they are members of, and provide a real-time custom experience of the game, with personalized offers and ads.
Peak Games, a mobile game developer, is using VoltDB with streaming data from Kinesis for real-time segmentation and A/B testing. This opens the opportunity for dynamic pricing to upgrade players or offering bonuses for 2X or 3X points. They can experiment with customized prices for a subset of players in real-time and perform A/B testing to optimize the offers and prices for in-game purchases.
Q: What are the most common hurdles you see affecting game development?
A: Helping clients understand what’s possible and how to use all of the tools to their advantage. In VoltDB, materialized views are a powerful way to aggregate tables of data in real time, group by segments, and summarize. Bringing logic to the data versus letting it reside on the application, which is inherently slow. Storing data in memory enables fast access for real-time changes in game dynamics and offers. What is possible today was unheard of a few years ago, and developers are still catching on to what today’s technology can help them build.
Q: What’s the future of game development?
A: Games that are personalized for the individual playing — knowing how they like to play, the history of what they’ve done — anything to increase engagement. Real-time intelligence supplements the intelligence in the game. Intelligence is built into the platform to inform AR and VR.
Q: What are your biggest concerns around game development today?
A: I’d like to see more games return to the intuitive gameplay of the classic arcade experience, where you knew how to play a game within seconds of watching someone play it, even if it still might take many hours to master. As we leverage technology to make games more engaging, it’s important not to make them too addictive. Let kids have fun within limits. Give parents the ability to set limits.
Q: What skills do developers need to develop successful games?
A: There are new languages, libraries, frameworks, especially with AR, VR, graphics, physics, and gameplay. Find technologies that leverage familiar languages or tools so you can gain a basic understanding of them and build on your knowledge over time. In building a data analytics platform for a game, it helps to know the basics of Linux and server-side languages such as Java, which have not changed in an unrecognizable way over the years, so you can avoid constantly re-learning things and become more proficient. That is one reason why new SQL databases are a good choice. Data engineers and analysts can leverage them as long as they know SQL. Bring the power of what you have learned before and build upon it.
Q: What have I failed to ask that you think we need to consider with regards to game development?
A: Fantasy sports is a segment of gaming that’s growing quickly. These are very engaging and very data-driven. Millions of people refresh their screen every time a player scores. End users want to see where they stand versus their friends and versus the entire player universe. Ultimately, this will generate more data and require more bandwidth than other games.
Opinions expressed by DZone contributors are their own.