Eight Common Software Architecture Design Patterns and Their Pros and Cons

Architecture is the skeleton of software, analogous to the human skeleton supporting muscles and nerves. This article shares eight common architecture design patterns, their typical designs, scenarios, advantages and disadvantages.

1. Single‑Database Single‑Application Pattern

Simple structure with one database, one business application layer, and a back‑office system. Suitable for prototypes or small‑scale products. Advantages: simple, fast development. Disadvantages: poor performance, limited scalability, not suitable for large production.

2. Content Distribution Pattern

Uses CDN and cloud storage (OSS) to deliver static resources close to users. Typical flow: upload → OSS → URL stored in DB → client fetches via intelligent DNS to nearest server. Advantages: fast resource download, reduces backend storage pressure. Disadvantages: cost, consistency and update latency.

3. Query Separation Pattern

Separates read and write by using master‑slave databases and introduces Elasticsearch for full‑text search. Improves read performance and reduces DB load. Advantages: high read throughput, flexible search. Disadvantages: data latency, consistency challenges.

4. Microservice Pattern

Decomposes a monolithic system into independent services, each with its own DB, cache, and ES, communicating via RPC or MQ. Solves single‑point failures, scaling, and team ownership issues. Advantages: high performance, strong scalability, suitable for medium‑to‑large companies. Disadvantages: increased complexity, operational overhead.

5. Multi‑Level Cache Pattern

Adds caching at client, API gateway, and backend layers to absorb read traffic. Improves response time and reduces backend load. Advantages: handles massive read requests. Disadvantages: cache invalidation and potential avalanche when caches miss.

6. Sharding (Database Partitioning) Pattern

Splits large tables across multiple databases or tables (horizontal/vertical sharding) to alleviate pressure on a single DB. Requires a routing layer to direct reads/writes. Advantages: reduces single‑table pressure. Disadvantages: distributed transaction complexity, extensive code changes.

7. Elastic Scaling Pattern

Provides automatic horizontal scaling of instances based on traffic spikes, often using VM or Docker resource pools. Eliminates manual over‑provisioning. Advantages: elastic resource usage, cost‑effective. Disadvantages: requires highly scalable architecture and robust monitoring.

8. Multi‑Datacenter (Multi‑Region) Pattern

Deploys services in multiple data centers to achieve high availability and low latency for global users. Handles data synchronization, routing, and failover. Advantages: high availability, high performance, multi‑region active‑active. Disadvantages: data consistency, synchronization, routing complexity.

Overall, the article emphasizes choosing the right pattern based on business needs, traffic characteristics, and team capabilities, and warns against over‑fragmentation.