Reengineering an adtech platform for efficient ad serving
About the client
Our client is a video ad platform that enables the world’s top marketers and media brands to reach consumers across TV, desktop and mobile devices through premium experiences, programmatic buying and performance-driven campaigns. It is a global partner of choice for ad publishers, agencies and advertisers.
Java, C, RIBS, Python, ROR, Flex, Angular JS, Jetty, Tomcat, Kafka, Netezza, Vertica, Redshift, MySQL
With growing digital demand, the ad ecosystem witnessed evolutions at a dynamic pace. Marketers and agencies needed access to more video inventory spots at better prices and with an easy buying process.On the other hand, publishers pursued a way to capture the full value of their inventory. To cater this opportunity, the client required a lean data collection and reporting infrastructure, demand/inventory forecaster and an upgrade to their existing UI. Special attention was required to deliver solutions for their open source driven backend.
We built ribified versions of MySQL library, SMTP client and ribs context swapping. Resolved performance issues by introducing lean objects using DTO patterns. To facilitate quick response times, we tweaked the JVM settings and introduced JDBC connection pooling over Netezza, this resulted in 20% reduction of the processing time.
The demand forecaster handled data inflow of almost 100GB/hour and still delivering sub second response times. Porting of core business logics, happened in record time ensuring business continuity. These made the platform highly competitive and ensured a funding of almost $45 Mn. Over its period of existence, the platform served close to 3.13 billion ads. It eventually got acquired by AOL in 2013 for a valuation of $405 Mn and got merged with its One platform.
We’re open for businessLet’s get started