| name | system-planning-assistant |
| description | High-level system architecture planning with component diagrams, scalability analysis, technology trade-offs, cost estimation, migration strategies, and technical specification documents |
Purpose
Assist with strategic system design decisions, analyzing trade-offs, defining architectures, and creating technical specifications for complex systems.
When to Use
Use this skill when the task involves:
- Designing new system architecture from requirements
- Evaluating technology choices and trade-offs
- Planning system migrations or refactoring
- Creating technical specification documents
- Estimating infrastructure costs
- Analyzing scalability and reliability requirements
Constraints
- Follow Dependency Rule — dependencies point inward to domain
- Keep domain logic independent of frameworks and infrastructure
- Treat microservices as deployment strategy, not architecture
- Apply Domain-Driven Design for complex business domains
- Document explicit trade-offs for every major decision
- Include non-functional requirements (latency, throughput, availability)
Execution Steps
- Gather requirements — Functional and non-functional (SLAs, scale targets)
- Identify domains — Bounded contexts and their relationships
- Design components — Services, databases, messaging, caching layers
- Create diagrams — C4 model (context, container, component)
- Analyze trade-offs — Document pros/cons of key decisions
- Estimate resources — Team size, timeline, infrastructure costs
- Define milestones — Phased delivery plan with risk mitigation
Examples
Input
Design architecture for a real-time bidding platform handling 100k requests/second
Output
# Real-Time Bidding Platform Architecture
## Requirements
- **Throughput**: 100,000 requests/second peak
- **Latency**: P99 < 50ms for bid decisions
- **Availability**: 99.95% SLA
## High-Level Architecture
```mermaid
graph TB
subgraph Edge
CDN[CloudFront CDN]
LB[Application Load Balancer]
end
subgraph Compute
BidService[Bid Service<br/>Auto-scaled EKS]
AuctionEngine[Auction Engine<br/>In-memory processing]
end
subgraph Data
Redis[(Redis Cluster<br/>Campaign cache)]
Kafka[Kafka<br/>Event streaming]
TimescaleDB[(TimescaleDB<br/>Analytics)]
end
CDN --> LB --> BidService
BidService --> Redis
BidService --> AuctionEngine
AuctionEngine --> Kafka
Kafka --> TimescaleDB
```
Key Decisions
| Decision | Choice | Trade-off |
|---|
| Caching | Redis Cluster | +Low latency, -Eventual consistency |
| Database | TimescaleDB | +Time-series optimized, -Less ecosystem |
| Messaging | Kafka | +Durability, -Operational complexity |
Cost Estimate (Monthly)
- EKS (50 nodes): ~$15,000
- Redis Cluster: ~$3,000
- Kafka (MSK): ~$2,500
- Data transfer: ~$5,000
- Total: ~$25,500/month
## Related Skills
- `terraform-aws-builder` — Implement infrastructure
- `database-schema-designer` — Design data models
- `kafka-stream-designer` — Event streaming design
