Elevator, Floor, Request classes in LLD - Scalability & System Analysis

Scalability Analysis - Elevator, Floor, Request classes

Growth Table: Elevator System Scaling

Users (People)	Requests per Minute	Elevators Needed	System Complexity	Latency (Wait Time)
100	50	1	Simple: Single elevator, direct request handling	Low (seconds)
10,000	5,000	10-20	Moderate: Multiple elevators, request queueing, scheduling	Moderate (tens of seconds)
1,000,000	500,000	100+	Complex: Distributed control, load balancing, fault tolerance	Variable, optimized by scheduling algorithms
100,000,000	50,000,000	Thousands	Very complex: Multi-building coordination, advanced predictive scheduling	Challenging, requires AI and real-time analytics

First Bottleneck

At small scale (100 users), the bottleneck is the elevator's mechanical speed and door operation time.

At medium scale (10,000 users), the bottleneck is the request processing and scheduling logic in the control system, as many requests compete for limited elevators.

At large scale (1 million+ users), the bottleneck shifts to communication and coordination between multiple elevator controllers and the central system, plus data processing delays.

Scaling Solutions

• Horizontal Scaling: Add more elevators to serve more requests concurrently.
• Load Balancing: Distribute requests evenly among elevators to avoid congestion.
• Caching: Cache frequent floor requests or patterns to optimize scheduling.
• Sharding: Divide floors or buildings into zones, each managed by separate controllers.
• Predictive Scheduling: Use historical data to anticipate demand and pre-position elevators.
• Fault Tolerance: Implement fallback mechanisms if an elevator or controller fails.

Back-of-Envelope Cost Analysis

• Requests per second at 10,000 users: ~83 (5,000 requests/min ÷ 60)
• Each elevator can handle ~5-10 requests per minute depending on speed and stops.
• Storage for request logs: Minimal, a few KB per request, scaling to GBs at large scale.
• Network bandwidth: Low per request, mostly control signals; scales with number of elevators and controllers.
• CPU: Scheduling algorithms run in milliseconds; more elevators increase CPU needs linearly.

Interview Tip

Start by defining the scale and key components: elevators, floors, requests.

Discuss how requests flow from users to elevators and how scheduling works.

Identify bottlenecks at each scale and propose targeted solutions.

Use real-world analogies like traffic lights or taxi dispatch to explain scheduling.

Always mention trade-offs: cost vs. latency vs. complexity.

Self Check Question

Your elevator control system handles 1000 requests per minute. Traffic grows 10x. What do you do first?

Answer: Add more elevators (horizontal scaling) and improve scheduling algorithms to handle increased requests efficiently.

Key Result

Elevator systems scale by adding more elevators and improving scheduling; the first bottleneck shifts from mechanical limits to control logic and communication as user count grows.