Understanding Analysis Modules
Understanding Analysis Modules
Architecto's analysis modules help you evaluate your architecture against multiple critical dimensions before deployment. Each module focuses on a specific aspect: cost, scalability, security, performance, or compliance.
Analysis Overview
Think of analysis modules like a peer review with specialists:
- Finance Auditor → Cost Estimator
- Performance Engineer → Scalability Analyzer, Performance Simulator
- Security Architect → Threat Modeler, Security Posture
- Operations Lead → Drift Detector
- Compliance Officer → Compliance Checker
Together, they give you confidence that your architecture is production-ready.
Cost Estimator
Question: "How much will this cost?"
The Cost Estimator models your monthly cloud bill at different scale points.
How It Works
- You input your architecture and target scale (1k → 1M users)
- Architecto calculates compute, storage, database, and data transfer costs
- It models how costs change as you scale
- Identifies cost inflection points — Where costs suddenly jump
- Suggests optimization strategies
Understanding Cost Curves
Your cost typically follows patterns:
Cost
^
| /
| /
| /
| /_____ Cost knee
| /
|/
+-----> UsersThe "cost knee" is where your architecture becomes financially unsustainable. For example:
- At 10k users: $200/month ✓ Healthy
- At 50k users: $500/month ✓ Still okay
- At 100k users: $2,000/month ⚠️ Cost knee
- At 200k users: $5,000/month ✗ Unsustainable
Cost Reduction Strategies
Common optimizations the Cost Estimator suggests:
| Strategy | Example | Impact |
|---|---|---|
| Reserved Instances | 1-year commitment on EC2 | 20-40% savings |
| Caching | Add Redis for hot data | 30-50% DB cost reduction |
| Autoscaling | Scale down at night | 10-20% compute savings |
| Data Transfer | Use CDN for static assets | 20-40% data transfer savings |
| Database Optimization | Use RDS read replicas, connection pooling | 15-30% DB costs |
| Spot Instances | For non-critical workloads | 70% compute savings |
Scalability Analyzer
Question: "How many users can this handle? Where's the bottleneck?"
The Scalability Analyzer uses queuing theory and stress-testing models to predict how your architecture handles load.
Key Metrics
| Metric | Definition | Example |
|---|---|---|
| Throughput | Requests/second the system can handle | 5,000 req/s |
| Latency | Response time for average request | 200ms |
| Bottleneck | Component that limits throughput | RDS primary database |
| Scaling Limit | Max users before quality degrades | 100,000 concurrent |
Identifying Bottlenecks
Different components bottleneck at different scales:
10k users: API Server (add more servers)
50k users: Database (add read replicas, caching)
100k users: Network bandwidth (add CDN, optimize queries)
1M users: Distributed systems complexity (redesign for microservices)The Scalability Analyzer pinpoints exactly which component becomes your constraint.
Scaling Recommendations
Once bottlenecks are identified:
-
Add Read Replicas (databases)
- Readers go to replicas, writers to primary
- Scales read throughput linearly
- Typical: 3-5 replicas handles 20-50k users
-
Horizontal Scaling (servers)
- Add more instances behind load balancer
- Typical: Each instance handles 500-2,000 concurrent users
-
Caching Layer (Redis/Memcached)
- Cache frequently-accessed data
- Reduces database load by 50-80%
- Typical: Cache layer handles 10-100x more traffic
-
Sharding (databases)
- Split data by key (user ID, region, date)
- Each shard handles subset of data
- Enables scaling to millions of users
-
Connection Pooling (databases)
- Reuse database connections
- Reduces connection overhead
- Typical: 2-3x throughput improvement
Threat Modeler
Question: "What could go wrong? How do we prevent attacks?"
The Threat Modeler uses STRIDE methodology to identify security threats and suggest mitigations.
STRIDE Framework
| Threat | Definition | Example |
|---|---|---|
| Spoofing | Attacker pretends to be someone else | Fake user login, forged API requests |
| Tampering | Attacker modifies data in transit or at rest | Man-in-the-middle attack, database corruption |
| Repudiation | Attacker denies performing an action | Denies deleting data, denies making transfer |
| Information Disclosure | Attacker reads data they shouldn't | Exposure of API keys, customer data leak |
| Denial of Service | Attacker prevents legitimate users from accessing service | DDoS, resource exhaustion |
| Elevation of Privilege | Attacker gains higher-level access than allowed | SQL injection to read all data, privilege escalation |
Example Threat Analysis
For your e-commerce architecture:
Spoofing Threat:
- Attacker impersonates another user
- Accesses their orders and payment methods
- Mitigation: MFA on login, OAuth 2.0, session tokens
Tampering Threat:
- Attacker modifies order price during checkout
- Changes $100 order to $1
- Mitigation: Cryptographic signatures, server-side validation, encrypted communication
Information Disclosure:
- Attacker reads database backup
- Steals customer credit cards and personal data
- Mitigation: Encrypt database at-rest (AWS KMS), encrypt backups, access controls
Denial of Service:
- Attacker floods API with requests
- Legitimate users get HTTP 503 errors
- Mitigation: Rate limiting, CloudFront DDoS protection, WAF, auto-scaling
Security Best Practices
Architecto recommends:
- Authentication — MFA, OAuth 2.0, passkeys
- Encryption — TLS 1.3 for transit, AES-256 for data at rest
- Authorization — Least-privilege IAM, role-based access control
- Network Security — VPC isolation, security groups, NACLs
- Data Protection — Tokenization of PII, masking in logs
- Monitoring — CloudTrail for audit logs, CloudWatch for anomalies
Architecture Diff
Question: "What changed between versions? What's the impact?"
The Architecture Diff shows semantic differences between two versions of your architecture.
What It Compares
- Added Components — New services, databases, etc.
- Removed Components — Deprecated services
- Modified Components — Configuration changes (instance type, scaling, etc.)
- Connection Changes — New/removed data flows
Example Diff
Version 1:
- Single RDS database (no replicas)
- Single EC2 instance
- No caching layer
- Direct database access from app
Version 2:
+ Added 2 RDS read replicas
+ Added autoscaling group (3-10 instances)
+ Added Redis cache layer
+ Added connection pooling
~ Changed database instance type (t3.medium → r5.large)Impact Analysis
The diff explains impacts:
- Cost Impact: +$500/month (for additional replicas and cache)
- Scalability Impact: +400% (can now handle 5x more users)
- Availability Impact: +99.9% (multi-AZ failover now configured)
- Complexity Impact: Medium (added cache and replicas to manage)
Performance Simulator
Question: "How will latency and throughput perform under load?"
(Coming soon: Full performance simulation with latency modeling)
The Performance Simulator predicts how each component behaves under different loads.
Compliance Checker
Question: "Does this meet compliance requirements?"
(Coming soon: SOC2, HIPAA, PCI DSS, GDPR validation)
The Compliance Checker validates your architecture against standards.
Drift Detector
Question: "Has our infrastructure drifted from the designed architecture?"
(Coming soon: Continuous monitoring for configuration drift)
Running Complete Analysis
All-in-One Review
- Click "Analyze All" (coming soon)
- Architecto runs all applicable modules:
- Cost estimate (1k, 10k, 100k, 1M users)
- Scalability bottleneck analysis
- Complete STRIDE threat review
- Compliance check
- Performance simulation
- Generates comprehensive Architecture Review Report:
- Executive summary (score 1-10)
- Detailed findings from each module
- Critical issues requiring attention
- Recommended next steps
- Cost/benefit of each recommendation
Tracking Analysis Over Time
Save multiple analysis runs to track:
- How architecture improves
- Cost trends
- Scaling readiness improvements
- Security posture improvements
Next Steps
- Create Your First Architecture → First Architecture
- Learn Best Practices → Best Practices
- Explore All Features → Features Guide
Key Takeaway:
Analysis modules aren't just reports—they're your architecture review board. Use them to catch problems early, optimize costs, and build confidence before deployment.