10 Billion a Day, 100 Milliseconds Per: Monitoring Real-Time Bidding at AdRoll

This talk will discuss AdRoll's mission-critical, real-time bidding system which competes for ad spaces globally while end-users' webpages load. This talk will focus on how we address monitoring and performance tracking with respect to scaling our "thundering herd" type load in a latency sensitive, high-concurrency environment. I will discuss the challenges of achieving detailed insight into a complex Erlang system and how we've addressed some of these, mitigated others. 

Talk objectives:

With this talk I hope to give some insight into a mission-critical erlang system running in an unusual latency and concurrency sensitive environment. I hope to disseminate some of the unique but generally applicable lessons learned from this environment and engage the community with an interesting anecdote or two. 

Target audience:

* Developers working in large-scale real-time environments
* Developers working in performance critical environments
* Developers working with or considering mission-critical erlang deployments
* "big data" aficionados
* Developers deploying erlang in Amazon's EC2 environment

Slides
Video

About Brian

Brian is a seven-year Erlang hacker, having gotten into it at University and his interests run to the large-scale distributed side of things. He's worked with Erlang professionally for four years, currently at AdRoll where he's a developer on the RTB (real-time bidding) team and previously at Rackspace, where he was a developer on the FireEngine project (previously discussed at Erlang Factory 2012).

Twitter: @bltroutwine

GitHub: blt 

Back to conference page