Project Details

Key Results

• •High quality media output, with lossless constant compression
• •High quality content recommendation system, each user has different preference, from top, down to toe, and the content delivery system will do just that
• •High Concurency streaming. Backed by lossless constant compressing using in-house rendering engine (powered by FFMPEG)
• •Current business model does not apply intrusive ads like popup, banner etc, it prioritize the user comfort.
• •Main income stream will come from Subscription based model, we give the user the ability to choose whether they wanted to get the ads, in exchange of accessing premium content
• •Neat & Tidy Video player developed in house (powered by customized Vidstack)
• •Supports HLS Streaming by default, and will fallback to HTML5 Video
• •Can Rival youtube for content delivery speed with limited server capability

Challenge

Multimedia is expensive, infrastructure wise, you need massive storage & server to process all of the content and delivering it to end-user conveniently. Handling concurrency, storage, content pre&post processing. a mature, complex engineering is required to deliver and execute it right. It's not just a CRUD App, it involved lots of Infrastructure from Distributed Remote Rendering server, Cloud Storage, CDN, and custom personalized content recommendation system. In which, most of them developed in-house. For this type of project, lots of manpower is needed to handle multiple aspect of the App.

Solution

Crafted most of the paid & vendor tool based in-house, this in turn cut down alot of cost, in expense of time. Distributed Remote Rendering is acting as microservice that pre&post process the uploaded content by admin. It applies range of parameter set during content initialization. Ranging from Watermarking, Thumbnail Generation, Thumbnail Sprites (Youtube like, when you move the slider to the left and right, it displays a tiny preview thumbnail). Preview Generation, Metadata Customization, Annotations, Parameter, and lots of parameter for the content recommendation system.