Notice

Paul Kiernan (pak79@cornell.edu) and Gabe Gralla (gig3@cornell.edu) are setting up a team for this project. They interned at Moat over the summer and are looking for anyone excited to design and tune a large scale, production database

Moat: Internet Advertising Analytics

Client

Max Parke, Moat
max.parke@moat.com

Moat

Moat is a NYC based startup in the Internet advertising analytics industry. Their main product measures ad effectiveness by analyzing users’ mouse hovers and creating heatmaps based on them (see http://www.moat.com/pages/analytics for examples).

The Project

A huge amount of data is collected from live ads, which requires efficient, reliable, and scalable systems to analyze it. The rate at which Moat is expanding requires a more efficient and reliable data store. Our project aims to develop a scalable, queryable, high availability, and redundant key-value data store, to record mouse movement and other metrics at the user level. We will likely build upon Redis, a database which performs writes in memory before they are later persisted to disk. A large portion of the project will be tuning the system to meet benchmarks set by the client.