Ongame Bets Safe on Columnar Database
Commentary by Mattias Andersson, Business Analyst, Ongame
Background
Ongame Network is the world’s largest B2B poker network, with more than 20 million registered users. Headquartered in Gibraltar, the organization provides the infrastructure that supports many of Europe’s leading online gaming and poker sites, including Bwin, Betfair, Unibet, Winga, and Bet 24. In addition to hosting online games, Ongame Network provides services to help partners maximize revenue generation, ranging from promotions and tournament development to loyalty schemes. Founded in 1999, Ongame currently operates in more than 25 markets, has an international network licensed in Gibraltar, and has regional networks in France and Italy.
The Challenge
Central to Ongame’s success in identifying revenue opportunities for its partners is the ability to capture and analyze data from multiple online poker games. To achieve this, Ongame implemented a MySQL database that worked well for a number of years. With the increasing popularity of online gambling, the business expanded rapidly. This resulted in a larger user base and growing demand for new gaming options and tournaments with greater diversity and complexity. With larger volumes of data needing analysis, it was apparent that the MySQL solution would struggle to process and analyze the resulting volumes of data.
“Looking back, we had probably reached the point where we had too much data for MySQL to be effective,” said Mattias Andersson, business analyst at Ongame. “With around 35,000 concurrent users, the database tables started to fill up rapidly. We had around 4.5 billion records in single database tables, so it was increasingly difficult to run queries and get responses in a reasonable timescale. A single query could take up to a day and sometimes it just wasn’t possible to ask the question—we didn’t have the functionality. As a result, we found it increasingly difficult to provide the detailed analysis that we wanted. For example, when doing analysis on player behavior, we could only use aggregates of a day played instead of looking at all player hands.”
Solution and Results
It was obvious that Ongame needed to implement a new database that could handle the existing and projected data volumes. After reviewing a number of options, Ongame selected the Infobright® Enterprise Edition database.
“With MySQL, we could only store details from one poker table for a single day. Now, we can store that data for a year―and still access it whenever it’s needed.” —Mattias Andersson, Business Analyst, Ongame
“We implemented Infobright’s solution quickly and found it very easy to set up, primarily because it doesn’t use indexes or other manual tuning to speed query performance,” Andersson said. “This meant we didn’t have to think about the specific details of the reports and queries we would want to run—and that saved us considerable amounts of time during the setup phase. Additionally, because of Infobright’s very high data compression, we were able to store far more data than we’d originally anticipated.”
Ongame’s query times are now fewer than 20 minutes, compared with more than a day with the previous solution. Faster response times, combined with the ability to store and analyze more data over longer periods, has allowed Ongame to give its partners the detailed information they need.
“The information we can now provide gives our partners the flexibility to develop their offerings to meet their individual revenue targets in a way that simply wasn’t previously possible,” said Andersson. “With MySQL, we could only store details from one poker table for a single day. Now, we can store that data for a year―and still access it whenever it’s needed. This level of detail gives partners the information they need. Then they can decide the best way forward, whether that’s changing rate structures and the ways rake is calculated, or altering bonus payouts and the level of earnings per player.”
This article originally appeared in the issue of .