Adjust: Scaling a Platform for Real-time Fraud Detection without Breaking the Bank
Adjust’s business is verifying billions of advertising impressions, policing any fraudulent activity at scale. They maintain all of the data on their clients’ target user devices and their actions – unlike their competitors which discard data after 90 days. Setting up a global, clustering system with transcontinental fibre to manage 100TB of data was not trivial. Data correctness and low latency were important criteria.
In this session, learn:
- Why they decided to move from NoSQL incumbent Redis to Aerospike
- Why bare metal works so much better than the cloud
- How they’ve scaled while still policing fraud at speed
- Their architectural view to manage the data volume, upgrades, backups and interruptions
Presenter: Robert Abraham – Adjust, Director of Engineering, Adjust. Robert oversees Adjust’s entire engineering department. He works to scale the company’s infrastructure, leads co-ordination with the Product team, and manages Adjust’s efforts to provide clients with a solution that gives them exactly what they need.
Note: This session was presented at Big Data London 2018