HIVE In Production at LivingSocial and Accumulo, a New Database From the NSA

Relational Data in a MapReduceWorld

Bryce Nyggen, Data Engineer

Web applications, and especially e-commerce ones, typically structure their data models for rapid insertion and lookup of individual people, catalog items, purchases, and so on. Hadoop and Hive will let you more or less seamlessly transition data from your RDBMS of choice to MapReduce, but to do analysis at scale, your data model and data infrastructure need to change as well. This talk will go over how we at LivingSocial effectively use Hadoop / Hive / HBase for analysis of RDBMS-based data, from the data model, to reporting infrastructure, to organizational concerns.

An Introduction to Accumulo and MapReduce over NoSQL

Aaron Cordova, Software Engineer, The Interllective Inc.

MapReduce and NoSQL databases are among the exciting recent breakthroughs in Big Data architectures that are designed for different purposes. This talk discusses the intersection of these two technologies and how MapReduce and NoSQL databases can work together to achieve even greater efficiency. We discuss these techniques in two particular instances: MapReduce over MongoDB - a popular NoSQL document database and Accumulo, a newly released BigTable implementation with fine-grained access controls and server-side computation from the NSA.

WAN Latency Optimization for Hadoop and Other Distributed Applications

Cancelled due to speaker availability.

Agenda:

6:00 PM - 6:30 PM - Refreshments and Networking

6:30 PM - 7:00 PM - First Speaker

7:00 PM - 7:15 PM - Break

7:15 PM - 7:45 PM - Second Speaker