# SRE & Platform Engineering Introductory Guide (For Executives)
**Understanding "Digital Infrastructure Investment" for Non-Engineers**

---

## Why You Should Read This Document

If you've ever felt:
- "The system is working, so why does the infrastructure team need budget?"
- "If we hire more developers, development should speed up, but why doesn't it?"
- "Can't we just respond when incidents happen?"

This document is for you.

**Time required: 10 minutes**
**After reading:** You'll be able to make investment decisions about the "invisible infrastructure" supporting digital services.

---

## 1. Premise: Digital Services Are "Cities," Not "Buildings"

### Common Misconception

"Once you build a system, it just keeps running with maintenance."

→ **This is incorrect.**

### Reality

Digital services:
- Grow continuously every day (user increase, feature additions)
- Must constantly adapt to changing environments (technology evolution, security threats)
- Must run 24/7/365 without rest

**In other words, digital services are closer to "cities."**
Just as cities need police, fire departments, water systems, and road maintenance, digital services need "invisible infrastructure."

---

## 2. What Is SRE: The "Police and Fire Services" of Digital Cities

### SRE = Site Reliability Engineering

**In one sentence:**
"A team that performs prevention and response to keep services running stably"

### Similarities with Police and Fire Services

| Police/Fire | SRE |
|---|---|
| Respond when fires or crimes occur | Respond when system incidents occur |
| Prevent through disaster drills, patrols | Prevent through monitoring systems, incident prediction |
| Inspect fire equipment | Check system health |
| Install security cameras, street lights | Build monitoring infrastructure, alerts |
| Prepare for large-scale disasters | Prepare for traffic surges |

### What SRE Does

**① Prevention activities (70%)**
- Build system monitoring infrastructure
- Design architecture that doesn't cause incidents
- Capacity planning ("Need server scaling in 3 months," etc.)
- Reduce human error through automation

**② Response activities (30%)**
- Emergency response when incidents occur
- Root cause analysis and recurrence prevention
- Post-mortems (retrospectives)

### Common Misconception

❌ "People who get called when problems occur"
⭕ "People who create mechanisms to prevent problems"

---

## 3. What Is Platform Engineering: The "Roads and Water Systems" of Digital Cities

### Platform Engineering = Infrastructure for Developer Efficiency

**In one sentence:**
"A team that creates a 'self-service infrastructure' where all developers can safely deploy themselves"

### Similarities with City Infrastructure

| City Infrastructure | Platform Engineering |
|---|---|
| Water: turn faucet, water comes out | Deploy: press button, goes to production |
| Roads: anyone can reach destinations | CI/CD: anyone can release safely |
| Power grid: electricity reaches all homes | Monitoring infrastructure: monitoring covers all services |
| Sewers: invisible but essential | Log infrastructure: invisible but essential for analysis |

### Why Necessary: How Developers Spend Time

**Without Platform Engineering:**

Developers manually:
- Spin up servers (30 minutes)
- Write deploy configuration (1 hour)
- Add monitoring settings (1 hour)
- Wait for someone's approval (hours to days)

**With Platform Engineering:**

Developers with one button:
- Standardized environment auto-starts (3 minutes)
- Automatic deployment (5 minutes)
- Monitoring auto-configured
- No approval needed (with guardrails)

### Business Impact

**Example: 50-person development team**

- 2 hours saved per deploy
- 20 deploys/week × 50 weeks = 2,000 hours saved annually
- Engineer hourly rate 5,000 yen × 2,000 hours = **10 million yen annual labor cost reduction**
- Platform construction cost 5 million yen → **Investment recovered in 6 months**

---

## 4. Why Investment Is Necessary: Visualizing "Invisible Losses"

### Costs of Not Investing

#### Case 1: Major Incident

- E-commerce site down for 2 hours
- Hourly revenue 5 million yen × 2 hours = **10 million yen opportunity loss**
- Loss of customer trust (hard to quantify, but the largest loss)

**Could have been prevented with 5 million yen annual SRE investment.**

#### Case 2: Slow Development Speed

- Each deploy takes half a day
- Can only release once a week
- Competitors release daily → **Lose on market speed**

**With 10 million yen annual Platform Engineering investment, can release daily.**

#### Case 3: Excellent Engineer Attrition

- "Too much manual work, can't do what I want"
- Engineer hiring/training cost: 5-10 million yen per person
- **2-3 people leave annually = 15-30 million yen loss**

### Investment Mindset: See It as "Insurance"

Companies pay millions of yen annually for fire insurance. They don't say "fire hasn't happened, so it's wasteful."

**Same thinking:**
- Expected loss: 30 million yen annually (incident risk)
- Investment: 5 million yen (SRE)
- **If investment < expected loss, it's rational**

---

## 5. How Much to Invest: Three Guidelines

### ① Engineer Ratio

**Guideline: 1 SRE per 10-20 developers**

- 50 developers → 3-5 SRE
- 100 developers → 5-10 SRE

※Varies with service scale and complexity

### ② Revenue Ratio

**Guideline: 10-20% of IT budget for operations/maintenance (including SRE)**

- Annual IT budget 100 million yen → 10-20 million yen

### ③ Calculate from Downtime Cost

**Formula:**
1. Calculate revenue lost per hour of downtime
2. Decide annual acceptable downtime (e.g., 99.9% = 8.76 hours)
3. Calculate potential annual loss
4. Invest 10-20% of that

**Example:**
- Downtime loss per hour: 5 million yen
- Annual acceptable downtime: 8.76 hours
- Potential loss: 43.8 million yen
- **Investment: 5-10 million yen**

---

## 6. Common Misconceptions and Truths

### Misconception 1: "It's working now, so it's fine"

**Truth:** Systems constantly degrade (technical debt, capacity limits). Even if working now, we don't know about 3 months from now.

### Misconception 2: "Just respond when incidents happen"

**Truth:** Damage from 1 major incident (revenue loss + trust loss) > several years of preventive investment

### Misconception 3: "Hiring more developers speeds things up"

**Truth:** Without infrastructure, it actually slows down (manual work queues, person-dependence, increased incidents).

### Misconception 4: "Cloud means no operations needed"

**Truth:** Cloud is a "tool" that requires SRE and Platform Engineering to use effectively.

### Misconception 5: "SRE doesn't contribute to development"

**Truth:** On the foundation SRE creates, developers can develop with confidence. Without foundation, you can't build a house.

---

## 7. What Happens Without Investment: Negative Spiral

```
Budget cuts
   ↓
Staff shortage, tool shortage
   ↓
Increased incidents, slower development
   ↓
"Why is it so unstable?"
   ↓
Further budget cuts (vicious cycle)
```

**Same structure with police/fire:**
Budget cuts → Reduced patrols → Increased crime → "What are police doing?" → Further budget cuts

---

## 8. Executive Decision Checklist

If you answer "No" to many of these questions, consider investing in SRE/Platform Engineering:

### SRE Related

- [ ] Zero service-impacting incidents in the past year
- [ ] System to respond within 30 minutes when incidents occur
- [ ] Monitoring infrastructure in place to detect problems proactively
- [ ] Can predict capacity 3 months ahead
- [ ] All engineers understand incident response documentation

### Platform Engineering Related

- [ ] Developers can deploy multiple times daily
- [ ] Can launch new services within 1 day
- [ ] Deploy operations fully automated
- [ ] Developers don't feel stressed by "wait times"
- [ ] Engineer attrition rate below industry average

---

## 9. Next Actions: Where to Start

### Step 1: Assess Current State (1 month)

- Calculate incident history and losses over past year
- Interview developers (what takes up their time)
- Measure deploy frequency and lead time

### Step 2: Start Small (3 months)

- Hire 1-2 SRE or transfer from within
- Pilot Platform Engineering initiatives in one team
- Measure effectiveness (deploy time, incident count, developer satisfaction)

### Step 3: Gradually Expand (6+ months)

- If effective, increase budget and roll out company-wide
- Regularly measure ROI and report at executive meetings

---

## 10. Summary: Digital Infrastructure Is "Investment," Not "Cost"

### What to Understand

1. **Digital services are "cities"**
   Just as police, fire, roads are needed, so are SRE and Platform Engineering

2. **"Invisible losses" are real**
   Revenue loss from incidents, slower development speed, talent attrition

3. **Preventive investment has high ROI**
   Loss from 1 incident > several years of preventive investment

4. **Industry standard: "1 SRE per 10-20 developers"**
   Below this risks falling into negative spiral

5. **Investment decisions are "risk management"**
   Like fire insurance: not betting it won't happen, but preparing for when it does

### Executive Perspective

**Not "Should we invest in SRE and Platform Engineering?"
but "How much investment is needed to support business growth?"**

If digital services are your business core, investment in that infrastructure is **not optional but mandatory**.

---

**For questions about this document or to discuss specific investment plans, please consult your CTO or VPoE.**