Strategy pattern in LLD - Scalability & System Analysis

Scalability Analysis - Strategy pattern

Growth Table: Strategy Pattern Usage

Users/Requests	What Changes?
100 users	Few strategies implemented; simple context switching; low memory and CPU usage.
10,000 users	More strategies added; increased context objects; moderate CPU usage; need for efficient strategy selection.
1,000,000 users	High concurrency; many strategy instances; potential CPU bottleneck; need caching or pooling of strategies.
100,000,000 users	Massive scale; strategy instantiation overhead critical; must optimize strategy reuse; consider distributed context handling.

First Bottleneck

The first bottleneck is CPU and memory usage due to frequent creation and switching of strategy objects under high load. As user requests grow, the overhead of instantiating and managing many strategy instances can slow down the system.

Scaling Solutions

• Object Pooling: Reuse strategy instances instead of creating new ones each time.
• Caching: Cache results of strategy computations when possible to avoid repeated work.
• Horizontal Scaling: Add more servers to distribute the load of strategy execution.
• Lazy Initialization: Instantiate strategies only when needed to save memory.
• Asynchronous Processing: Offload heavy strategy computations to background workers.

Back-of-Envelope Cost Analysis

Assuming each user request triggers one strategy execution:

• At 1,000 QPS, CPU usage rises due to object creation and method calls.
• Memory usage grows with number of strategy instances; pooling reduces this.
• Network bandwidth impact is minimal as strategies run in-process.
• Storage is not significantly affected unless strategies cache data persistently.

Interview Tip

When discussing scalability of the Strategy pattern, start by explaining how strategy objects are created and used. Then identify the overhead of instantiation and switching at scale. Propose concrete solutions like pooling and caching. Finally, mention horizontal scaling and asynchronous processing as ways to handle very high load.

Self Check

Your database handles 1000 QPS. Traffic grows 10x. What do you do first?

Answer: Since the database is the bottleneck, first add read replicas or caching layers to reduce load. For the Strategy pattern, similarly, first optimize strategy instance reuse and caching before scaling servers.

Key Result

The Strategy pattern scales well at low to moderate load but faces CPU and memory bottlenecks at high concurrency due to frequent strategy instantiation; pooling, caching, and horizontal scaling are key solutions.

Practice

1. What is the main purpose of the Strategy pattern in system design?

easy

A. To restrict object creation to a single instance

B. To allow selecting an algorithm's behavior at runtime without changing the client code

C. To define a fixed sequence of steps for an algorithm

D. To create a single global instance of a class

Solution

1. Step 1: Understand the Strategy pattern goal

   The Strategy pattern is designed to let you swap algorithms or behaviors dynamically without changing the client code.

2. Step 2: Compare options with pattern purpose

   To allow selecting an algorithm's behavior at runtime without changing the client code correctly states this purpose. Options A and B describe Singleton pattern, and C describes Template Method pattern.

3. Final Answer:

   To allow selecting an algorithm's behavior at runtime without changing the client code -> Option B

4. Quick Check:

   Strategy pattern = runtime algorithm selection [OK]

Hint: Strategy pattern = choose behavior at runtime [OK]

Common Mistakes:

• Confusing Strategy with Singleton pattern
• Thinking Strategy fixes algorithm steps
• Assuming Strategy creates single instances

2. Which of the following is the correct way to define a Strategy interface in a typical object-oriented language?

easy

A. class Strategy { void execute(); }

B. def Strategy(): pass

C. interface Strategy { void execute(); }

D. struct Strategy { void execute(); }

Solution

1. Step 1: Identify the correct syntax for interface definition

   In many object-oriented languages like Java or C#, interface keyword is used to define a Strategy interface with method signatures.

2. Step 2: Evaluate options

   interface Strategy { void execute(); } uses interface with method execute(), which is correct. class Strategy { void execute(); } defines a class, not an interface. def Strategy(): pass is Python syntax but incomplete. struct Strategy { void execute(); } uses struct which is not typical for interfaces.

3. Final Answer:

   interface Strategy { void execute(); } -> Option C

4. Quick Check:

   Strategy interface = interface with method [OK]

Hint: Strategy interface uses 'interface' keyword in OOP [OK]

Common Mistakes:

• Using class instead of interface for Strategy
• Confusing struct with interface
• Using incomplete or wrong language syntax

3. Given the following code snippet using the Strategy pattern, what will be the output?

class Context:
    def __init__(self, strategy):
        self.strategy = strategy
    def execute(self):
        return self.strategy.do_action()

class StrategyA:
    def do_action(self):
        return 'Action A'

class StrategyB:
    def do_action(self):
        return 'Action B'

context = Context(StrategyB())
print(context.execute())

medium

A. Action B

B. Action A

C. Error: StrategyB has no method do_action

D. None

Solution

1. Step 1: Trace object creation and method calls

   The Context is created with StrategyB() instance. Calling context.execute() calls StrategyB.do_action().

2. Step 2: Check StrategyB.do_action() return value

   StrategyB.do_action() returns the string 'Action B', so print(context.execute()) outputs 'Action B'.

3. Final Answer:

   Action B -> Option A

4. Quick Check:

   Context with StrategyB = 'Action B' output [OK]

Hint: Context calls strategy's method, output matches chosen strategy [OK]

Common Mistakes:

• Assuming default strategy is StrategyA
• Thinking method do_action is missing
• Confusing class and instance usage

4. Identify the error in the following Strategy pattern implementation:

class Context:
    def __init__(self, strategy):
        self.strategy = strategy
    def execute(self):
        return self.strategy.action()

class StrategyA:
    def do_action(self):
        return 'Action A'

context = Context(StrategyA())
print(context.execute())

medium

A. Context calls a method 'action' which does not exist in StrategyA

B. StrategyA class is missing the constructor

C. Context should not store strategy as an instance variable

D. No error, code runs correctly

Solution

1. Step 1: Compare method names between Context and StrategyA

   Context calls self.strategy.action(), but StrategyA defines do_action(), not action().

2. Step 2: Identify mismatch causing error

   This mismatch causes an AttributeError at runtime because action() is undefined in StrategyA.

3. Final Answer:

   Context calls a method 'action' which does not exist in StrategyA -> Option A

4. Quick Check:

   Method name mismatch = runtime error [OK]

Hint: Check method names match between context and strategy [OK]

Common Mistakes:

• Ignoring method name mismatch
• Assuming missing constructor causes error
• Thinking strategy should not be stored in context

5. You are designing a payment system that supports multiple payment methods (credit card, PayPal, cryptocurrency). How would applying the Strategy pattern improve your system design?

hard

A. It requires hardcoding all payment methods inside a single class

B. It forces all payment methods to share the same implementation details

C. It prevents runtime selection of payment methods

D. It allows adding new payment methods without changing existing code by defining each as a separate strategy

Solution

1. Step 1: Understand Strategy pattern benefits in payment methods

   Strategy pattern lets you define each payment method as a separate strategy class implementing a common interface.

2. Step 2: Analyze how this affects system design

   This design allows adding new payment methods easily without modifying existing code, and lets the system select payment method at runtime.

3. Final Answer:

   It allows adding new payment methods without changing existing code by defining each as a separate strategy -> Option D

4. Quick Check:

   Strategy pattern = easy extension and runtime choice [OK]

Hint: Strategy pattern enables easy addition and runtime choice [OK]

Common Mistakes:

• Thinking Strategy forces shared implementation
• Believing all methods must be hardcoded together
• Assuming runtime selection is not possible