Play Framework | AWS | NoSQL | Partners | PostgreSQL | ArchitectureWelcoming Nina DayAugust 10, 2018
Neil Chaudhuri (He/Him)
Neil Chaudhuri (He/Him)
Vidya is proud to be working with Nina Day, a casting and creative agency based in New York City serving a wide array of clients worldwide. Nina Day finds the best talent--models, actors, musicians, and many others--for ads for Fortune 500 companies, TV and movies, and special campaigns like the United Nations He for She campaign promoting gender equality worldwide.
As you might imagine, a world class casting company needs a talent database, and we have built one for Nina Day. It's a web application deployed with SSL/TLS that provides a host of features including the following:
Secure self-entry for talent
High-performance upload of talent assets--images, videos, resumes, etc.--to the cloud and subsequent download and streaming
The ability to create presentations for clients by defining categories, organizing talent within categories, and organizing files by drag-and-drop for each talent
Fast, dynamic query capability for finding the right talent for mass export to presentations and/or broadcast E-mail
The ability for clients to view presentations (where they have permissions) and select the talent they prefer after viewing images and videos and generating PDFs for the presentation, a category, or individual talent
The ability to view selections by user along with common selections made by multiple users and generate PDFs of those as well
The Nina Day talent database began as a Ruby on Rails web application with primarily a JQuery and Bootstrap frontend and a PostgreSQL and Amazon S3 backend with CloudFront CDN for worldwide distribution of images and videos.
Over time though, we found that this architecture couldn't scale as needed--particularly with regard to long-running background work that takes place with tasks like image and video processing and PDF generation. Enter Play Framework in Scala. We migrated the talent database to Play with a MongoDB backend for a more flexible schema--not to mention that a document-based database is a natural fit for the data model.
Play and Scala are built for performance, and they have made the database so much faster. We also use akka-s3 from the Alpakka project for asynchronous streaming to S3. This has proven very helpful for reducing the memory footprint and speeding things up as well.
Some things stayed the same like S3 and CloudFront--albeit with a new key structure--and the JQuery/Bootstrap front end. That's changing somewhat as we are currently migrating the user interface to a more modern stack with Webpack and React. We spent some time assessing the suitability of React, Elm, and Vue for the talent database UI, and as compelling as we find Elm in particular, we decided on React because of the established community and consequently the availability of components that easily replicate existing functionality. Besides, we have also integrated TypeScript, RxJS, and Immutable to provide a fair bit of the type safety, immutability, and functional reactive programming that make Elm so cool.
We are lucky to be working with Nina Day to use so many powerful, modern technologies to build software that is clean, fast, and secure for our client and theirs worldwide.