MemSQL, the leader in distributed in-memory database technology, recently announced that global digital-media holding company CPXi recently deployed MemSQL v3.0 in order to utilize all of the company’s data—250 billion rows of real-time and historical data—for its real-time bidding operations. CPXi provides multi-screen messaging that leverages display, social, mobile and video advertising at scale and serves billions of managed impressions daily. Extract, transform and load (ETL) is expensive and required CPXi to have extra machines and storage to run this time-intensive process. By switching from Hadoop to MemSQL, CPXi eliminated ETL processes and also cut 50 percent of its Amazon Elastic Compute Cloud (EC2) instances and 50TB of storage, which led to hundreds of thousands of dollars in annual savings.
As we tested multiple vendors, we came to see that we don’t just have Big Data, we have Huge Data, which brings new problems and complications,” said Mike Zacharski, chief operating officer at CPXi. “With its real-time queries, cost-effective data accessibility and reliability, MemSQL provided the right solution to meet our unique business requirements—including our large data sets—so that we can operate more effectively and efficiently and push innovation forward within the highly competitive ad tech space.”
Digital media clients are increasingly demanding real-time bidding (also known as programmatic ad buying), a dynamic auction process that ingests and analyzes billions of data points in real time to produce fast and accurate bids and more targeted ads. But many media companies are struggling to implement real-time bidding because of the sheer volume of data and velocity of processing that it requires.
Before MemSQL, CPXi was one of those companies: loading data from its tiered architecture into its analysis tools was an expensive, cumbersome process that could take 12 to 24 hours, which meant that data aged and became less relevant before it could be analyzed. MemSQL eliminated that problem: CPXi now does a front-line ingest into row store and then a real-time transfer of the data to the column store for analysis. This new consolidated, tiered storage architecture with a unified SQL interface eliminates CPXi’s ETL process and simplifies the complexity of its database infrastructure. The architecture also helps CPXi to scale up staffing by exposing a familiar ANSI SQL interface so that new employees can come up to speed swiftly.
Performance and reliability are crucial in the ad tech space, and CPXi needed the ability to scale out and analyze data quickly in order to provide the most accurate results for their customers,” said Eric Frenkiel, co-founder and CEO of MemSQL. “We find our customers have Big Data challenges that require solutions that provide immediate business value. CPXi is a testament to the power of our database, and we’re proud to be its valued partner in helping it to better serve customers.”
Sign up for the free insideBIGDATA newsletter.