2nd/10_Wiki/Topics/DevOps_and_Security/엔터프라이즈 소프트웨어 시스템 설계.md

---
id: wiki-2026-0508-엔터프라이즈-소프트웨어-시스템-설계
title: 엔터프라이즈 소프트웨어 시스템 설계
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Enterprise Software Design, Enterprise Architecture]
duplicate_of: none
source_trust_level: A
confidence_score: 0.9
verification_status: applied
tags: [enterprise, architecture, systems-design, scalability]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: java
  framework: spring-boot
---

# 엔터프라이즈 소프트웨어 시스템 설계

## 매 한 줄
> **"매 large-scale, mission-critical, long-lived"**. 매 enterprise software 의 multi-team / multi-decade scale 의 system 의 design — 매 reliability, security, compliance, integration 의 first-class concern. 매 2026 의 cloud-native + 매 legacy mainframe 의 hybrid 의 reality.

## 매 핵심

### 매 Enterprise 의 specific concerns
- **Scale**: 매 thousands of concurrent users, TB+ data.
- **Longevity**: 매 10-30년 의 lifespan — 매 maintainability 의 critical.
- **Compliance**: 매 SOX, GDPR, HIPAA, PCI-DSS.
- **Integration**: 매 N legacy system 의 talk to.
- **Multi-team**: 매 Conway's law — 매 org structure 가 architecture 의 reflect.

### 매 Architecture layers (전통)
1. **Presentation** — UI / API.
2. **Application** — orchestration, use cases.
3. **Domain** — business logic.
4. **Infrastructure** — persistence, messaging, external.

### 매 2026 modern stack
- **Frontend**: React / Next.js, micro-frontends.
- **API**: GraphQL Federation, gRPC, REST.
- **Service**: Microservices (Java/Spring Boot, Go, Kotlin).
- **Data**: PostgreSQL, Kafka, Snowflake / Databricks.
- **Platform**: Kubernetes, Istio, ArgoCD.
- **Observability**: OpenTelemetry, Datadog, Grafana.

### 매 Cross-cutting concerns
- AuthN/AuthZ (OAuth2, OIDC, mTLS)
- Audit logging
- Distributed tracing
- Rate limiting, circuit breakers
- Multi-tenancy

## 💻 패턴

### 매 Hexagonal architecture
```java
// 매 Domain — 매 framework 의존 X
public class OrderService {
    private final OrderRepository repo;     // port
    private final PaymentGateway payment;   // port

    public Order place(PlaceOrderCommand cmd) {
        var order = Order.create(cmd);
        payment.charge(cmd.payment());
        return repo.save(order);
    }
}
```

### 매 CQRS + Event Sourcing
```java
// Command side
@Component
class OrderCommandHandler {
    public void handle(PlaceOrder cmd) {
        var events = OrderAggregate.place(cmd);
        eventStore.append(cmd.orderId(), events);
        bus.publish(events);
    }
}

// Query side
@Component
class OrderProjection {
    @KafkaListener(topics = "order-events")
    public void on(OrderPlaced e) { readModel.upsert(e); }
}
```

### 매 Saga (distributed transaction)
```java
@Saga
public class OrderSaga {
    @StartSaga
    public void on(OrderPlaced e) {
        send(new ReserveInventory(e.orderId()));
    }

    @SagaEventHandler
    public void on(InventoryReserved e) {
        send(new ChargePayment(e.orderId()));
    }

    @SagaEventHandler
    public void on(PaymentFailed e) {
        send(new ReleaseInventory(e.orderId()));
        end();
    }
}
```

### 매 Strangler fig (legacy migration)
```yaml
# 매 Gateway routing — 매 old → new 의 incremental
routes:
  - path: /api/orders/*
    backend: new-order-service  # 매 migrated
  - path: /api/billing/*
    backend: legacy-mainframe   # 매 still legacy
```

### 매 Multi-tenancy 의 patterns
```java
// 매 Row-level — 매 cheapest
@Entity
class Order {
    @Column UUID tenantId;
    // 매 every query 의 tenant filter
}

// 매 Schema-per-tenant — 매 isolation
DataSource ds = tenantRouter.resolve(currentTenant());

// 매 DB-per-tenant — 매 strongest, costliest
```

### 매 BFF (Backend for Frontend)
```yaml
services:
  bff-mobile:    # 매 mobile 최적화 payload
  bff-web:       # 매 web 최적화
  bff-partner:   # 매 B2B integration
  # 매 모두 매 동일 backend service 의 aggregate
```

## 매 결정 기준
| 상황 | Approach |
|---|---|
| 매 startup, ≤10 dev | 매 Modular monolith |
| 매 scale-up, multi-team | 매 Microservices + DDD |
| 매 legacy 의 modernization | 매 Strangler fig |
| 매 strict audit / replay | 매 Event sourcing |
| 매 read-heavy + complex write | 매 CQRS |

**기본값**: 매 modular monolith 부터 → 매 pain point 별 의 carve out.

## 🔗 Graph
- 부모: [[Distributed Systems]]
- 변형: [[Microservices]] · [[Event-Driven Architecture]] · [[Service-Oriented Architecture]]
- 응용: [[E-commerce_Platforms|E-commerce Platforms]]
- Adjacent: [[Domain-Driven Design]] · [[Hexagonal Architecture]]

## 🤖 LLM 활용
**언제**: 매 large-scale system design interview / RFC / ADR 작성 시.
**언제 X**: 매 small CRUD app — 매 over-engineering 의 risk.

## ❌ 안티패턴
- **Distributed monolith**: 매 microservice 의 모양 의 만 갖춘 의 tight coupling.
- **Big bang rewrite**: 매 legacy 의 한 번에 의 replace — 매 거의 fail.
- **Anemic domain**: 매 logic 의 service layer 의 만 — 매 domain object 의 의 data bag.
- **Shared database**: 매 microservice 간 의 DB 공유 → 매 hidden coupling.

## 🧪 검증 / 중복
- Verified — Martin Fowler, *Patterns of Enterprise Application Architecture*; Eric Evans, *Domain-Driven Design*; Sam Newman, *Building Microservices* (2nd ed).
- 신뢰도 A.

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — enterprise architecture patterns + 2026 stack |