Please log in to watch this conference skillscast.

In the BBC’s Content Distribution Services division, they build and maintain systems that expose content metadata to be consumed by a wide range of audience-facing products. Their current architecture for distributing tagging metadata consists mainly of two JVM-based read and write APIs feeding off a Triplestore. This single storage setup for both read and write operations imposes restrictions on performance and scalability.

Augustine will share with you about his current stream of work, which involves creating an event-driven write pipeline that generates materialised views of tagging metadata corresponding to the various read profiles thus improving the overall performance and scalability of the system. The new architecture comprises of small, single-purpose Javascript-based services, lambda functions, event stores, queues, streams and a number of specialised data stores.

Building a majority of the components in Javascript, you will discover how they:

• Built simple components that are easy and quick to deploy.
• Used Typescript for static type-checking.
• Adopted Functional Javascript techniques.
• Adopted input validation libraries like joi.
• Used docker-compose to orchestrate dependencies for acceptance testing.
• Took advantage of other tools and libraries like Bluebird, nsp, testdouble, etc.

YOU MAY ALSO LIKE:

• ProgNET London 2019 (in London on 11th - 13th September 2019)
• Advanced Python (in London on 27th - 28th June 2019)
• Beginner Steps into Data with Inspiring Women in Data Science (in London on 25th April 2019)