Lesson 1: Design a Video Streaming Service

Goal: Design a video sharing platform like YouTube or Netflix where users can upload and watch videos.

Requirements

Functional

• Upload videos.
• Watch videos (streaming).
• Support multiple resolutions (360p, 720p, 1080p).

Non-Functional

• Reliability: No buffering.
• Availability: Videos are always accessible.
• Scalability: Handle millions of concurrent viewers.

Core Design

1. Storage (Blob Store)

Videos are large binary files (BLOBs). Databases are bad for this.

• Object Storage: AWS S3, Google Cloud Storage.
• Metadata: Store title, description, and S3 URL in a SQL/NoSQL DB.

2. Processing (Transcoding)

Raw uploads are huge. We need to convert them into different formats and resolutions.

• Transcoding Service: Breaks video into chunks and encodes them (H.264, VP9).

3. Delivery (CDN)

Serving video from a single server is too slow for global users.

• Content Delivery Network (CDN): Caches video chunks in edge servers close to the user.

4. Adaptive Bitrate Streaming (HLS/DASH)

The player automatically switches quality based on the user’s internet speed.

---

🛠️ Sruja Perspective: Modeling Infrastructure

We can use Sruja’s deployment nodes to visualize the global distribution of content.

import { * } from 'sruja.ai/stdlib'


YouTube = system "Video Platform" {
    WebApp = container "Web App"
    API = container "API Server"

    Transcoder = container "Transcoding Service" {
        description "Converts raw video to HLS format"
        scale { min 50 }
    }

    S3 = database "Blob Storage" {
        description "Stores raw and processed video files"
    }

    MetadataDB = database "Metadata DB"

    WebApp -> API "HTTPS"
    API -> MetadataDB "Reads/Writes"
    API -> S3 "Uploads"
    API -> Transcoder "Triggers"
    Transcoder -> S3 "Reads/Writes"
    Transcoder -> MetadataDB "Updates status"
}

// Deployment View
deployment GlobalInfra "Global Infrastructure" {
    node OriginDC "Origin Data Center" {
        containerInstance WebApp
        containerInstance API
        containerInstance Transcoder
        containerInstance S3
    }

    node CDN "CDN (Edge Locations)" "Cloudflare / Akamai" {
        // Represents cached content
        node USEast "US-East Edge"
        node Europe "Europe Edge"
        node Asia "Asia Edge"
    }
}

User = person "Viewer"

// Streaming Flow
scenario WatchVideo "User watches a video" {
    User -> WebApp "Get Video Page"
    WebApp -> API "Get Metadata (Title, URL)"
    API -> MetadataDB "Query"
    API -> User "Return Video Manifest URL"
    User -> CDN "Request Video Chunk (1080p)"
    CDN -> User "Stream Chunk"
}

// Upload Flow
scenario UploadVideo "Creator uploads a video" {
    User -> YouTube.WebApp "Upload Raw Video"
    YouTube.WebApp -> YouTube.API "POST /upload"
    YouTube.API -> YouTube.S3 "Store Raw Video"
    YouTube.API -> YouTube.Transcoder "Trigger Transcoding Job"
    YouTube.Transcoder -> YouTube.S3 "Read Raw / Write HLS"
    YouTube.Transcoder -> YouTube.MetadataDB "Update Video Status"
}

// Data flow: Video transcoding pipeline
flow TranscodingPipeline "Video Transcoding Data Flow" {
    YouTube.S3 -> YouTube.Transcoder "Streams raw video chunks"
    YouTube.Transcoder -> YouTube.Transcoder "Encodes to HLS (360p, 720p, 1080p)"
    YouTube.Transcoder -> YouTube.S3 "Writes encoded chunks"
    YouTube.Transcoder -> YouTube.MetadataDB "Updates manifest URLs"
}

// Data flow: Video delivery pipeline
flow DeliveryPipeline "Video Delivery Data Flow" {
    YouTube.S3 -> CDN "Replicates video chunks to edge"
    CDN -> User "Streams chunks on demand"
    User -> CDN "Requests next chunk based on bandwidth"
    CDN -> YouTube.S3 "Cache miss: fetch from origin"
}

// Data flow: Analytics pipeline
flow AnalyticsPipeline "Video Analytics Data Flow" {
    YouTube.WebApp -> YouTube.API "Sends view events"
    YouTube.API -> YouTube.MetadataDB "Updates view count"
    YouTube.API -> YouTube.MetadataDB "Stores watch time"
    YouTube.MetadataDB -> YouTube.API "Aggregates analytics"
}

view index {
title "Complete Video Platform"
include *
}

// Data flow view: Focus on data pipelines
view dataflow {
title "Data Flow View"
include YouTube.Transcoder YouTube.S3 YouTube.MetadataDB
exclude YouTube.WebApp YouTube.API
description "Shows data transformation and storage flows"
}

// Processing view: Transcoding pipeline
view processing {
title "Processing Pipeline"
include YouTube.Transcoder YouTube.S3
exclude YouTube.WebApp YouTube.API YouTube.MetadataDB
description "Focuses on video processing components"
}