Non-Functional Requirements

Non-functional requirements define how well the system performs its functions. They describe the quality attributes of the system.

Core Quality Attributes

Non-Functional Requirements
├── Scalability
├── Availability
├── Reliability
├── Performance
├── Security
├── Maintainability
├── Portability
└── Usability

Quality Attributes Overview

Attribute	Definition	Example Metric
Scalability	Handle growing load	10K → 1M users
Availability	Uptime percentage	99.99% uptime
Reliability	Correct operation	< 0.01% error rate
Performance	Speed of response	< 200ms latency
Security	Protection from threats	No data breaches
Maintainability	Ease of modification	Changes in < 1 day

Non-Functional vs Functional

Functional: "User can upload a video"
Non-Functional: "Video upload completes within 5 seconds for files up to 100MB"

Functional: "User can search products"
Non-Functional: "Search results return within 500ms for 99% of queries"

The Illities

Many non-functional requirements end in "-ility":

• Scalability: Can grow
• Availability: Always accessible
• Reliability: Works correctly
• Maintainability: Easy to change
• Portability: Works on different platforms
• Extensibility: Easy to add features
• Testability: Easy to verify

Performance requirements specify how fast and efficiently the system must operate.

Key Performance Metrics

Latency
├── Response Time: Time for server to respond
├── Time to First Byte (TTFB): Time until first data received
└── Round Trip Time (RTT): Total time for request-response

Throughput
├── Requests Per Second (RPS)
├── Transactions Per Second (TPS)
└── Data Transfer Rate (MB/s)

Performance Targets

System Type	Latency Target	Throughput Target
Real-time Chat	< 100ms	10K messages/sec
E-commerce	< 500ms	1K orders/sec
Search Engine	< 200ms	100K queries/sec
Video Streaming	< 1s startup	1M streams

Performance Example: URL Shortener

Performance Requirements:
- URL creation: < 200ms (99th percentile)
- URL redirect: < 50ms (99th percentile)
- Analytics query: < 1s (95th percentile)
- System throughput: 100K redirects/sec

Measuring Performance

Client ──── 10ms ──── Network ──── 5ms ──── Server ──── 50ms ──── Database
           (Client)              (Network)            (Processing)           (Query)

Total Response Time = Client + Network + Server + Database
                    = 10 + 5 + 50 + 50 = 115ms

Performance Optimization Areas

1. Network: CDN, compression, HTTP/2
2. Server: Caching, connection pooling, async processing
3. Database: Indexing, query optimization, read replicas
4. Application: Algorithm optimization, lazy loading, pagination

Scalability requirements define how the system handles growing demand.

Types of Scalability

Scalability
├── Vertical Scaling (Scale Up)
│   └── Bigger machines
├── Horizontal Scaling (Scale Out)
│   └── More machines
└── Geographic Scaling
    └── More regions

Scalability Metrics

Metric	Current	Target
Users	10K	10M
Requests/sec	100	100K
Data Size	10GB	10TB
Concurrent Connections	1K	1M

Scalability Example: Chat System

Scalability Requirements:
- Support 10M concurrent users
- Handle 100K messages/sec peak
- Store 1B messages (10TB data)
- Scale to 100M users within 2 years

Design Implications:
- Horizontal scaling for stateless services
- Message queue for async processing
- Data partitioning by user ID
- CDN for static assets

When to Specify Scalability

Always specify in system design interviews:

1. Current scale: What's the expected starting load?
2. Growth rate: How fast will it grow?
3. Target scale: What's the maximum expected load?
4. Scaling timeline: How quickly do we need to scale?

Scalability vs Performance

Performance = How fast for current load
Scalability = How well it handles increased load

Good Performance + Poor Scalability = Fast now, slow later
Poor Performance + Good Scalability = Slow now, still slow later
Good Performance + Good Scalability = Fast now and later