RTB and Big Data - where Erlang and Hadoop meet

Philip Clarke

RTB (Real-time Bidding) is an on-the-fly auctioning system for delivering Ads. The key players that partake in such a solution are SSPs (Supply Side Platform) and DSPs (Demand Side Platform). At Adtech we have build an SSP solution that allows publishers to get better value for their advertisements. This solution however brings its own challenges where the key components need to support this on large scale. To be able to process billions of bid requests per day the realtime components are build in Erlang while the data part is based on Hadoop. This talk will give insights on how we use Erlang to build our RTBSelector fleet supporting horizontal scaling. In addition we’ll talk about the Big Data architecture that takes care about the downstream processing.


Philip has been working at AOL for 3 years as a Senior Erlang Developer.

Before that he worked at Ericsson developing software for mobile telephone systems in Plex and C++. This enabled Philip to become very familiar with the concurrent and fault tolerant problems that Erlang was designed to solve. 

It was only when Philip moved to AOL that he got the opportunity to develop highly concurrent soft real time systems using Erlang.

Back to conference page