Why database choice impacts architecture in HLD - Design It to Understand It

Design: Database Impact on System Architecture

In scope: database types and their architectural impact, data flow changes, scaling considerations. Out of scope: detailed database internals or vendor-specific features.

Functional Requirements

FR1: Explain how different database types affect system design decisions

FR2: Show how database choice influences scalability, latency, and data consistency

FR3: Demonstrate impact on data modeling and query patterns

FR4: Highlight effects on deployment and maintenance

Non-Functional Requirements

NFR1: Focus on common database types: relational, document, key-value, graph

NFR2: Consider typical scale: up to 100K concurrent users

NFR3: Latency target: p99 API response under 200ms

NFR4: Availability target: 99.9% uptime

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

Key Components

Database type (SQL, NoSQL variants)

Caching layers

Data access APIs

Replication and sharding mechanisms

Backup and recovery

Design Patterns

CQRS (Command Query Responsibility Segregation)

Event sourcing

Cache aside pattern

Sharding and partitioning

Polyglot persistence

Reference Architecture

Client
  |
  v
API Layer
  |
  v
+---------------------+
|  Database Layer      |
|  +----------------+ |
|  | Relational DB  | |
|  | Document DB    | |
|  | Key-Value DB   | |
|  | Graph DB       | |
|  +----------------+ |
+---------------------+
  |
  v
Cache Layer (optional)

Components

API Layer

REST/gRPC

Handles client requests and routes to appropriate database or cache

Relational Database

PostgreSQL/MySQL

Structured data with strong consistency and complex joins

Document Database

MongoDB/Couchbase

Flexible schema for semi-structured data, fast reads/writes

Key-Value Store

Redis/DynamoDB

Ultra-fast lookups for simple key-based data

Graph Database

Neo4j/JanusGraph

Efficient handling of complex relationships and traversals

Cache Layer

Redis/Memcached

Reduce database load and improve read latency

Request Flow

1. Client sends request to API Layer

2. API Layer determines data type and query complexity

3. For simple key lookups, API checks Cache Layer first

4. If cache miss, API queries appropriate database type

5. Database returns data to API Layer

6. API Layer returns response to client

7. Writes go directly to database with optional cache invalidation

Database Schema

Entities depend on database type: - Relational DB: Tables with normalized relations (1:N, N:N) - Document DB: Collections with nested documents - Key-Value DB: Simple key and value pairs - Graph DB: Nodes and edges representing entities and relationships Relationships and data models must align with chosen database capabilities

Scaling Discussion

Bottlenecks

Relational DB can become slow with complex joins at scale

Document DB may face consistency challenges with distributed writes

Key-Value stores limited by memory size and network latency

Graph DBs can struggle with very large or deeply connected data

Cache invalidation complexity increases with write volume

Solutions

Use read replicas and sharding for relational DB scaling

Implement eventual consistency and conflict resolution for document DB

Scale key-value stores horizontally and optimize network

Partition graph data and limit traversal depth

Use cache aside pattern and TTLs to manage cache freshness

Interview Tips

Time: Spend 10 minutes understanding requirements and clarifying data needs, 20 minutes discussing database types and architectural impact, 10 minutes on scaling and trade-offs, 5 minutes summarizing.

Explain how data shape and access patterns influence database choice

Discuss consistency vs availability trade-offs

Highlight how database choice affects latency and scalability

Mention caching and replication strategies

Show awareness of polyglot persistence for complex systems