Parking strategy pattern in LLD - System Design Exercise

Design: Parking Lot System with Strategy Pattern

Design focuses on the parking strategy pattern implementation and core parking operations. User authentication, payment processing, and physical hardware integration are out of scope.

Functional Requirements

FR1: Support multiple parking strategies (e.g., nearest spot, largest spot, electric vehicle spot).

FR2: Allow dynamic selection and switching of parking strategies at runtime.

FR3: Handle vehicle parking and unparking operations.

FR4: Track available and occupied parking spots.

FR5: Support different vehicle types (car, motorcycle, electric car).

Non-Functional Requirements

NFR1: System should handle up to 500 concurrent parking/unparking requests.

NFR2: Parking and unparking operations should complete within 200ms.

NFR3: System availability target is 99.9% uptime.

NFR4: Design should be extensible to add new parking strategies without major code changes.

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

Key Components

ParkingLot manager

ParkingSpot entity

Vehicle entity

Strategy interface and concrete strategy implementations

Spot availability tracker

Design Patterns

Strategy pattern for parking algorithms

Factory pattern to create strategy instances

Observer pattern to update spot availability

Singleton pattern for ParkingLot manager

Reference Architecture

  +-------------------+
  |   ParkingLot      |
  |  (Singleton)      |
  +---------+---------+
            |
            v
  +-------------------+        +-----------------------+
  | ParkingStrategy    |<-------| Concrete Strategies   |
  | (interface)       |        | - NearestSpotStrategy |
  +---------+---------+        | - LargestSpotStrategy |
            |                  | - ElectricSpotStrategy|
            v                  +-----------------------+
  +-------------------+
  | ParkingSpot       |
  +-------------------+
            ^
            |
  +-------------------+
  | Vehicle           |
  +-------------------+

Components

ParkingLot

Singleton class

Manages parking spots, vehicles, and delegates parking to selected strategy.

ParkingStrategy

Interface or abstract class

Defines method to find suitable parking spot for a vehicle.

NearestSpotStrategy

Concrete strategy class

Parks vehicle in the nearest available spot.

LargestSpotStrategy

Concrete strategy class

Parks vehicle in the largest available spot.

ElectricSpotStrategy

Concrete strategy class

Parks electric vehicles in spots with charging stations.

ParkingSpot

Entity class

Represents a parking spot with attributes like size, type, and availability.

Vehicle

Entity class

Represents a vehicle with type and size.

Request Flow

1. User requests to park a vehicle.

2. ParkingLot receives request and invokes current ParkingStrategy.

3. ParkingStrategy searches for suitable spot based on its algorithm.

4. If spot found, ParkingLot marks spot as occupied and assigns vehicle.

5. ParkingLot confirms parking success to user.

6. For unparking, ParkingLot frees the spot and updates availability.

Database Schema

Entities: - Vehicle(vehicle_id PK, type, size) - ParkingSpot(spot_id PK, size, type, is_occupied boolean) - ParkingAssignment(assignment_id PK, vehicle_id FK, spot_id FK, start_time, end_time nullable) Relationships: - One Vehicle can have zero or one active ParkingAssignment. - One ParkingSpot can have zero or one active ParkingAssignment. - ParkingAssignment links Vehicle and ParkingSpot with timestamps.

Scaling Discussion

Bottlenecks

ParkingLot manager becomes a bottleneck with many concurrent requests.

Searching for spots in large parking lots may be slow.

Updating spot availability under high concurrency can cause race conditions.

Solutions

Partition parking lot into zones with separate managers to reduce contention.

Use efficient data structures (e.g., priority queues, spatial indexes) for spot search.

Implement optimistic locking or distributed locks for spot updates.

Cache frequently accessed spot availability data in memory.

Interview Tips

Time: 10 minutes for requirements and clarifications, 15 minutes for design and patterns, 10 minutes for scaling and trade-offs, 10 minutes for Q&A.

Explain how strategy pattern enables flexible parking algorithms.

Discuss how to add new strategies without changing existing code.

Describe how concurrency and availability are handled.

Mention data model supporting parking assignments and spot tracking.

Highlight scalability approaches for large parking lots.

Practice

1. What is the main purpose of using the Parking Strategy Pattern in a parking lot system?

easy

A. To store vehicle details in a database

B. To manage payment processing for parking fees

C. To allow different algorithms for finding parking spots without changing the main system

D. To control the physical gates of the parking lot

Solution

1. Step 1: Understand the role of strategy pattern

   The strategy pattern lets you swap different algorithms easily without changing the main code.

2. Step 2: Apply this to parking system context

   In parking, it means you can change how spots are found without rewriting the whole system.

3. Final Answer:

   To allow different algorithms for finding parking spots without changing the main system -> Option C

4. Quick Check:

   Strategy pattern = flexible parking spot finding [OK]

Hint: Strategy pattern = flexible algorithm swapping [OK]

Common Mistakes:

• Confusing strategy pattern with data storage
• Thinking it controls hardware like gates
• Mixing it with payment processing logic

2. Which of the following is the correct way to define a parking strategy interface in a typical object-oriented design?

easy

A. class ParkingStrategy { findSpot(vehicle) {} }

B. function findSpot(vehicle) { return spot; }

C. var ParkingStrategy = new Object();

D. interface ParkingStrategy { findSpot(vehicle): Spot; }

Solution

1. Step 1: Identify interface syntax in OOP

   Interfaces define method signatures without implementation, e.g., interface ParkingStrategy { findSpot(vehicle): Spot; }.

2. Step 2: Compare options

   interface ParkingStrategy { findSpot(vehicle): Spot; } correctly uses interface with method signature. Others are class, function, or object, not interface.

3. Final Answer:

   interface ParkingStrategy { findSpot(vehicle): Spot; } -> Option D

4. Quick Check:

   Interface = method signature only [OK]

Hint: Interface defines method signatures only [OK]

Common Mistakes:

• Using class instead of interface for strategy
• Defining functions outside interface context
• Confusing object creation with interface definition

3. Given the following code snippet for two parking strategies, what will be the output when findSpot(vehicle) is called using NearestParkingStrategy?

class NearestParkingStrategy {
  findSpot(vehicle) {
    return "Nearest spot found";
  }
}
class RandomParkingStrategy {
  findSpot(vehicle) {
    return "Random spot found";
  }
}
const strategy = new NearestParkingStrategy();
console.log(strategy.findSpot('car'));

medium

A. "Nearest spot found"

B. "Random spot found"

C. undefined

D. Error: findSpot not defined

Solution

1. Step 1: Identify which strategy instance is used

   The code creates an instance of NearestParkingStrategy and calls findSpot.

2. Step 2: Check method output for that class

   NearestParkingStrategy.findSpot returns "Nearest spot found".

3. Final Answer:

   "Nearest spot found" -> Option A

4. Quick Check:

   Instance method output = "Nearest spot found" [OK]

Hint: Instance method output matches class used [OK]

Common Mistakes:

• Confusing which strategy instance is created
• Assuming random strategy is used
• Expecting undefined or error without reason

4. In the following code, what is the main issue that prevents the ParkingLot from using different parking strategies correctly?

class ParkingLot {
  constructor() {
    this.strategy = null;
  }
  setStrategy(strategy) {
    this.strategy = strategy;
  }
  park(vehicle) {
    return this.strategy.findSpot(vehicle);
  }
}
const lot = new ParkingLot();
lot.park('car');

medium

A. No strategy is set before calling park, causing an error

B. The park method should not call findSpot

C. The strategy property should be a list, not a single object

D. The constructor should initialize strategy with a default value

Solution

1. Step 1: Analyze object initialization and method calls

   The ParkingLot is created with strategy = null. No strategy is set before calling park.

2. Step 2: Understand consequence of null strategy

   Calling this.strategy.findSpot(vehicle) when strategy is null causes an error.

3. Final Answer:

   No strategy is set before calling park, causing an error -> Option A

4. Quick Check:

   Null strategy causes error on method call [OK]

Hint: Always set strategy before use [OK]

Common Mistakes:

• Thinking park method logic is wrong
• Assuming strategy should be a list
• Ignoring null initialization problem

5. You want to design a parking system that supports multiple vehicle types (car, bike, truck) and different parking strategies (nearest, random, reserved). Which design approach best uses the Parking Strategy Pattern to handle this complexity?

hard

A. Create separate parking lot classes for each vehicle type and hardcode the strategy inside each

B. Use a single ParkingLot class with a strategy interface; implement different strategies and select based on vehicle type at runtime

C. Store all vehicles in one list and assign spots randomly without strategy pattern

D. Use global variables to track vehicle types and apply strategies in procedural code

Solution

1. Step 1: Identify need for flexibility and scalability

   Supporting multiple vehicle types and strategies requires flexible design without code duplication.

2. Step 2: Apply strategy pattern with runtime selection

   Using one ParkingLot class with strategy interface allows swapping strategies dynamically based on vehicle type.

3. Step 3: Evaluate other options

   Create separate parking lot classes for each vehicle type and hardcode the strategy inside each duplicates code, C ignores strategy pattern, D uses poor global state management.

4. Final Answer:

   Use a single ParkingLot class with a strategy interface; implement different strategies and select based on vehicle type at runtime -> Option B

5. Quick Check:

   Strategy pattern + runtime selection = best design [OK]

Hint: Use one class + strategy interface + runtime choice [OK]

Common Mistakes:

• Duplicating classes per vehicle type
• Ignoring strategy pattern benefits
• Using global variables instead of OOP