Ben Christensen describes how the Netflix API evolved from a typical one-size-fits-all RESTful API designed to support public developers into a web service platform optimized to handle the diversity and variability of each device and user experience. The talk will also address the challenges involving operations, deployment, performance, fault-tolerance, and rate of innovation at massive scale.
Kyle Boon reviews 3 frameworks for building RESTful WS- Grails, Dropwizard and Ratpack-, comparing their code readability, maintainability, deployment, metrics collection, scalability and testability.
Brendan Eich surveys interesting developments in the Web platform, analysing emergent trends, and making some predictions.
Nick Kolegraff discusses common problems and architecture to support all the phases of data science and how to start a data science initiative, sharing lessons from Accenture, Best Buy, and Rackspace.
Dan Woods discusses using Spring Integration and design patterns to implement a message-driven architecture in Grails to allow for better modularity, scalability, and code reusability.
Jeremy Edberg discusses how Netflix designs their systems and deployment processes to help the service survive both catastrophic events like zone and regional outages and less catastrophic events like network latency and random instance death.
Blake Dournaee covers the often forgotten back-end architecture for mobile apps which should expose cross-platform APIs to mitigate some of the effects of mobile O/S fragmentation.
Steve Pember discusses creating Grails applications integrating message broker technologies, especially RabbitMQ, and applying SOA principles.
Paco Nathan reviews an example data analysis application written in Cascalog used for a recommender system based on City of Palo Alto Open Data.
Jeff Magnusson takes a deep dive into key services of Netflix’s “data platform as a service” architecture, including RESTful services that: provide comprehensive metadata management across data sources (Franklin); enable visualization and caching of results of Hadoop jobs (Sting); and visualize the execution plans produced by languages such as Pig and Hive (Lipstick).
The panelists discuss innovation at the enterprise level avoiding the business’ constant volatility.
Crista Lopes writes a program in multiple styles -monolithic/OOP/continuations/relational/Pub-Sub/Monads/AOP/Map-reduce- showing the value of using more than a style in large scale systems.