System Design

Case Study: Uber

Requirements

Functional

• Rider requests ride
• Match with nearby driver
• Real-time location tracking
• ETA calculation
• Pricing

Scale

• 100M active users
• 10M trips/day
• Real-time GPS updates

Core Challenge: Matching

Problem: Eng yaqin driver’ni topish (location-based search).

Geospatial Indexing

QuadTree

Divide map into quadrants
Each quadrant → 4 sub-quadrants
Leaf nodes → drivers in that area

class QuadTree {
  search(lat, lon, radius) {
    // Find quadrant
    // Return drivers in quadrant
  }
  
  update(driverId, lat, lon) {
    // Move driver to new quadrant
  }
}

Geohash

Location → String encoding
lat: 37.7749, lon: -122.4194
→ Geohash: "9q8yy"

Nearby locations have similar prefixes
9q8yy1, 9q8yy2, ... (all nearby)

SELECT * FROM drivers
WHERE geohash LIKE '9q8yy%'
AND status = 'available'
LIMIT 10;

Architecture

Rider App → API Gateway
          → Matching Service (finds drivers)
          → Location Service (QuadTree/Redis)
          → Driver App (WebSocket)

Real-time updates:
Driver → Location Service (GPS)
       → Update QuadTree
       → Broadcast to nearby riders

ETA Calculation

// Simple: Distance / average_speed
const distance = haversine(rider, destination);
const eta = distance / 50; // 50 km/h average

// Advanced: Historical traffic data
const eta = routingService.calculate(rider, destination, time_of_day);

Pricing (Surge)

// Demand > Supply → Surge pricing
const demand = activeRiders(area);
const supply = availableDrivers(area);

if (demand / supply > 2) {
  surgeMultiplier = 1.5x;
}

Scaling

10M trips/day:

• ~120 requests/sec
• Millions of GPS updates/sec

Strategies:

• Geosharding (city-based)
• Redis for hot data
• Kafka for GPS streams
• Cassandra for trip history

Xulosa

32 mavzu to’liq tayyor!

O’rgangan narsalar: 1-4: Asoslar (Kirish, Scalability, Tarmoq, Protokollar) 5-9: Kengayish (Load balancing, Scaling, Stateless, CDN, Proxy) 10-15: Ma’lumotlar (ACID, SQL/NoSQL, Indexes, Replication, Sharding, CAP) 16-19: Keshlash (Basics, Invalidation, Redis, Eventual) 20-23: Asinxron (Queues, Pub/Sub, Events, Microservices) 24-27: Qo’shimcha (Rate limit, Monitoring, Security, DR) 28-32: Case Studies (Interview, URL Shortener, Twitter, WhatsApp, Uber)

System Design masterlik sari!