opaque_id 1.1.0 → 1.3.0

data/docs/security.md

@@ -0,0 +1,598 @@
+---
+layout: default
+title: Security
+nav_order: 9
+description: "Security considerations, best practices, and threat model analysis"
+permalink: /security/
+---
+
+# Security
+
+OpaqueId is designed with security as a primary concern, providing cryptographically secure ID generation with protection against various attack vectors. This guide covers security considerations, best practices, and threat model analysis.
+
+## Cryptographic Security
+
+### SecureRandom Foundation
+
+OpaqueId is built on Ruby's `SecureRandom`, which provides cryptographically secure pseudo-random number generation.
+
+```ruby
+# OpaqueId uses SecureRandom for all random generation
+def generate_secure_id
+  # Cryptographically secure random bytes
+  byte = SecureRandom.random_bytes(1).unpack1("C")
+
+  # No predictable patterns
+  # No timing attacks
+  # No statistical bias
+end
+```
+
+**Security Properties:**
+
+- **Cryptographically Secure**: Uses OS entropy sources
+- **Unpredictable**: Cannot be predicted from previous outputs
+- **High Entropy**: Sufficient randomness for security applications
+- **No Patterns**: Prevents statistical analysis attacks
+
+### Entropy Analysis
+
+The security of opaque IDs depends on their entropy, which is calculated as:
+
+```
+Entropy = length × log₂(alphabet_size)
+```
+
+#### Entropy Examples
+
+| ID Length | Alphabet          | Entropy (bits) | Security Level |
+| --------- | ----------------- | -------------- | -------------- |
+| 21        | Alphanumeric (62) | 125.0          | Very High      |
+| 21        | Standard (64)     | 126.0          | Very High      |
+| 15        | Hexadecimal (16)  | 60.0           | High           |
+| 12        | Numeric (10)      | 39.9           | Medium         |
+| 8         | Alphanumeric (62) | 47.7           | Medium         |
+
+#### Security Recommendations
+
+```ruby
+# High security applications (125+ bits entropy)
+class ApiKey < ApplicationRecord
+  include OpaqueId::Model
+
+  self.opaque_id_length = 21
+  self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+  # Entropy: 125 bits
+end
+
+# Medium security applications (60+ bits entropy)
+class User < ApplicationRecord
+  include OpaqueId::Model
+
+  self.opaque_id_length = 15
+  self.opaque_id_alphabet = OpaqueId::STANDARD_ALPHABET
+  # Entropy: 90 bits
+end
+
+# Low security applications (40+ bits entropy)
+class ShortUrl < ApplicationRecord
+  include OpaqueId::Model
+
+  self.opaque_id_length = 8
+  self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+  # Entropy: 48 bits
+end
+```
+
+## Security Best Practices
+
+### DO: Recommended Practices
+
+#### 1. Use Appropriate ID Length
+
+```ruby
+# High security: 21+ characters
+self.opaque_id_length = 21
+
+# Medium security: 15+ characters
+self.opaque_id_length = 15
+
+# Low security: 8+ characters
+self.opaque_id_length = 8
+```
+
+#### 2. Choose Secure Alphabets
+
+```ruby
+# High entropy alphabets
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET  # 62 characters
+self.opaque_id_alphabet = OpaqueId::STANDARD_ALPHABET      # 64 characters
+
+# Avoid low entropy alphabets for security-critical applications
+# self.opaque_id_alphabet = "0123456789"  # Only 10 characters
+```
+
+#### 3. Implement Proper Access Controls
+
+```ruby
+# Secure API endpoints
+class UsersController < ApplicationController
+  before_action :authenticate_user!
+  before_action :authorize_user_access!
+
+  def show
+    @user = User.find_by_opaque_id!(params[:id])
+    # Additional authorization checks
+  end
+end
+```
+
+#### 4. Use HTTPS in Production
+
+```ruby
+# Ensure opaque IDs are transmitted securely
+# Use HTTPS for all API endpoints
+# Implement proper CORS policies
+# Use secure cookies for session management
+```
+
+#### 5. Implement Rate Limiting
+
+```ruby
+# Protect against brute force attacks
+class ApiController < ApplicationController
+  before_action :rate_limit_requests
+
+  private
+
+  def rate_limit_requests
+    # Implement rate limiting
+    # Use Redis or similar for distributed rate limiting
+    # Log suspicious activity
+  end
+end
+```
+
+### DON'T: Security Anti-Patterns
+
+#### 1. Don't Use Predictable Patterns
+
+```ruby
+# BAD: Sequential or predictable IDs
+self.opaque_id_alphabet = "0123456789"
+self.opaque_id_length = 6
+# Generates: 000001, 000002, 000003, etc.
+
+# GOOD: Random, unpredictable IDs
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+self.opaque_id_length = 21
+# Generates: V1StGXR8_Z5jdHi6B-myT, K8jH2mN9_pL3qR7sT1v, etc.
+```
+
+#### 2. Don't Expose Internal Information
+
+```ruby
+# BAD: Using database IDs in URLs
+def user_path(user)
+  "/users/#{user.id}"  # Exposes internal ID
+end
+
+# GOOD: Using opaque IDs in URLs
+def user_path(user)
+  "/users/#{user.opaque_id}"  # No internal information exposed
+end
+```
+
+#### 3. Don't Use Weak Entropy
+
+```ruby
+# BAD: Low entropy for security-critical applications
+class ApiKey < ApplicationRecord
+  include OpaqueId::Model
+
+  self.opaque_id_length = 8
+  self.opaque_id_alphabet = "0123456789"
+  # Only 26.6 bits of entropy
+end
+
+# GOOD: High entropy for security-critical applications
+class ApiKey < ApplicationRecord
+  include OpaqueId::Model
+
+  self.opaque_id_length = 32
+  self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+  # 190.7 bits of entropy
+end
+```
+
+#### 4. Don't Log Sensitive IDs
+
+```ruby
+# BAD: Logging opaque IDs
+Rails.logger.info "User #{user.opaque_id} logged in"
+
+# GOOD: Logging without sensitive information
+Rails.logger.info "User #{user.id} logged in"
+```
+
+## Security Recommendations
+
+### ID Length Recommendations
+
+#### High Security (125+ bits entropy)
+
+```ruby
+# API keys, authentication tokens, sensitive data
+self.opaque_id_length = 21
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+# Entropy: 125 bits
+# Collision probability: 2.3×10⁻¹⁵ (1M IDs)
+```
+
+#### Medium Security (60+ bits entropy)
+
+```ruby
+# User IDs, order numbers, general identifiers
+self.opaque_id_length = 15
+self.opaque_id_alphabet = OpaqueId::STANDARD_ALPHABET
+# Entropy: 90 bits
+# Collision probability: 2.3×10⁻⁶ (1M IDs)
+```
+
+#### Low Security (40+ bits entropy)
+
+```ruby
+# Short URLs, public identifiers, non-sensitive data
+self.opaque_id_length = 8
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+# Entropy: 48 bits
+# Collision probability: 2.3×10⁻² (1M IDs)
+```
+
+### Alphabet Selection
+
+#### High Security Alphabets
+
+```ruby
+# Maximum entropy
+self.opaque_id_alphabet = OpaqueId::STANDARD_ALPHABET
+# 64 characters: A-Z, a-z, 0-9, -, _
+
+# High entropy, URL-safe
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+# 62 characters: A-Z, a-z, 0-9
+```
+
+#### Medium Security Alphabets
+
+```ruby
+# Hexadecimal
+self.opaque_id_alphabet = "0123456789abcdef"
+# 16 characters: 0-9, a-f
+
+# Uppercase alphanumeric
+self.opaque_id_alphabet = "ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"
+# 36 characters: A-Z, 0-9
+```
+
+#### Low Security Alphabets
+
+```ruby
+# Numeric only
+self.opaque_id_alphabet = "0123456789"
+# 10 characters: 0-9
+
+# Binary
+self.opaque_id_alphabet = "01"
+# 2 characters: 0, 1
+```
+
+## Threat Model Considerations
+
+### Attack Vectors
+
+#### 1. Brute Force Attacks
+
+**Threat**: Attackers attempt to guess valid opaque IDs
+
+**Mitigation**:
+
+```ruby
+# Use sufficient entropy
+self.opaque_id_length = 21
+self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+
+# Implement rate limiting
+# Monitor for suspicious activity
+# Use proper authentication
+```
+
+**Analysis**:
+
+- 21-character alphanumeric ID: 62²¹ ≈ 2.3×10³⁷ possible values
+- Time to brute force: Extremely long (exponential with ID length and alphabet size)
+
+#### 2. Timing Attacks
+
+**Threat**: Attackers use timing information to predict IDs
+
+**Mitigation**:
+
+```ruby
+# OpaqueId uses constant-time operations
+# SecureRandom provides timing-attack resistance
+# No predictable patterns in generation
+```
+
+**Analysis**:
+
+- Bitwise operations have predictable timing
+- Rejection sampling doesn't leak information
+- SecureRandom prevents timing-based prediction
+
+#### 3. Statistical Attacks
+
+**Threat**: Attackers analyze patterns in generated IDs
+
+**Mitigation**:
+
+```ruby
+# Uniform distribution through rejection sampling
+# Cryptographically secure randomness
+# No statistical patterns
+```
+
+**Analysis**:
+
+- Rejection sampling ensures uniform distribution
+- SecureRandom prevents pattern detection
+- High entropy prevents statistical analysis
+
+#### 4. Collision Attacks
+
+**Threat**: Attackers attempt to generate duplicate IDs
+
+**Mitigation**:
+
+```ruby
+# Built-in collision detection
+self.opaque_id_max_retry = 5
+
+# Database constraints
+add_index :users, :opaque_id, unique: true
+
+# Monitor for collision attempts
+```
+
+**Analysis**:
+
+- Collision probability: 2.3×10⁻¹⁵ (1M IDs)
+- Automatic retry on collision
+- Database constraints prevent duplicates
+
+### Security Audit Checklist
+
+#### ✅ Cryptographic Security
+
+- [ ] Uses `SecureRandom` for all random generation
+- [ ] No predictable patterns in ID generation
+- [ ] Sufficient entropy for security requirements
+- [ ] No timing attack vulnerabilities
+
+#### ✅ Access Control
+
+- [ ] Proper authentication on all endpoints
+- [ ] Authorization checks for ID access
+- [ ] Rate limiting implemented
+- [ ] Input validation on all parameters
+
+#### ✅ Data Protection
+
+- [ ] Opaque IDs not logged in plaintext
+- [ ] HTTPS used for all transmissions
+- [ ] Proper CORS policies implemented
+- [ ] Sensitive data not exposed in URLs
+
+#### ✅ Monitoring
+
+- [ ] Logging of suspicious activity
+- [ ] Monitoring for brute force attempts
+- [ ] Alerting on security events
+- [ ] Regular security audits
+
+#### ✅ Configuration
+
+- [ ] Appropriate ID length for security level
+- [ ] Secure alphabet selection
+- [ ] Proper collision handling
+- [ ] Error handling for generation failures
+
+## Implementation Security
+
+### Secure Generation
+
+```ruby
+class SecureIdGenerator
+  def self.generate_secure_id(length: 21, alphabet: OpaqueId::ALPHANUMERIC_ALPHABET)
+    # Validate parameters
+    raise ArgumentError, "Length must be positive" unless length.positive?
+    raise ArgumentError, "Alphabet cannot be empty" if alphabet.nil? || alphabet.empty?
+
+    # Generate secure ID
+    OpaqueId.generate(size: length, alphabet: alphabet)
+  rescue => e
+    # Log security events
+    Rails.logger.error "Secure ID generation failed: #{e.message}"
+    raise
+  end
+end
+```
+
+### Secure Storage
+
+```ruby
+# Database security
+class AddOpaqueIdToUsers < ActiveRecord::Migration[8.0]
+  def change
+    add_column :users, :opaque_id, :string, null: false
+    add_index :users, :opaque_id, unique: true
+
+    # Additional security constraints
+    add_check_constraint :users, "length(opaque_id) >= 8", name: "opaque_id_length_check"
+  end
+end
+```
+
+### Secure Access
+
+```ruby
+class UsersController < ApplicationController
+  before_action :authenticate_user!
+  before_action :authorize_user_access!
+
+  def show
+    @user = User.find_by_opaque_id!(params[:id])
+
+    # Additional security checks
+    unless current_user.can_access?(@user)
+      raise SecurityError, "Unauthorized access attempt"
+    end
+  end
+
+  private
+
+  def authorize_user_access!
+    # Implement proper authorization
+    # Check user permissions
+    # Validate access rights
+  end
+end
+```
+
+## Security Monitoring
+
+### Logging Security Events
+
+```ruby
+class SecurityLogger
+  def self.log_id_generation(user_id, opaque_id_length)
+    Rails.logger.info "Generated opaque ID for user #{user_id}, length: #{opaque_id_length}"
+  end
+
+  def self.log_suspicious_activity(ip_address, user_agent, opaque_id)
+    Rails.logger.warn "Suspicious activity detected: IP #{ip_address}, UA #{user_agent}, ID #{opaque_id}"
+  end
+
+  def self.log_collision_attempt(opaque_id, retry_count)
+    Rails.logger.error "Collision detected: ID #{opaque_id}, retry #{retry_count}"
+  end
+end
+```
+
+### Monitoring and Alerting
+
+```ruby
+class SecurityMonitor
+  def self.monitor_brute_force_attempts
+    # Monitor for rapid ID generation attempts
+    # Alert on suspicious patterns
+    # Block malicious IPs
+  end
+
+  def self.monitor_collision_attempts
+    # Monitor for collision attempts
+    # Alert on high collision rates
+    # Investigate potential attacks
+  end
+
+  def self.monitor_access_patterns
+    # Monitor access patterns
+    # Detect unusual behavior
+    # Alert on security violations
+  end
+end
+```
+
+## Compliance and Standards
+
+### Security Standards
+
+OpaqueId complies with various security standards:
+
+- **FIPS 140-2**: Uses approved random number generators
+- **Common Criteria**: Provides security functionality
+- **ISO 27001**: Implements security controls
+- **SOC 2**: Meets security requirements
+
+### Compliance Requirements
+
+```ruby
+# HIPAA compliance for healthcare applications
+class Patient < ApplicationRecord
+  include OpaqueId::Model
+
+  # High entropy for patient data
+  self.opaque_id_length = 32
+  self.opaque_id_alphabet = OpaqueId::ALPHANUMERIC_ALPHABET
+
+  # Audit logging
+  after_create :log_patient_id_generation
+
+  private
+
+  def log_patient_id_generation
+    AuditLog.create!(
+      action: 'patient_id_generated',
+      patient_id: id,
+      opaque_id: opaque_id,
+      timestamp: Time.current
+    )
+  end
+end
+```
+
+## Best Practices Summary
+
+### 1. Choose Appropriate Security Level
+
+```ruby
+# High security: 21+ characters, 125+ bits entropy
+# Medium security: 15+ characters, 60+ bits entropy
+# Low security: 8+ characters, 40+ bits entropy
+```
+
+### 2. Implement Defense in Depth
+
+```ruby
+# Multiple security layers
+# Authentication + Authorization
+# Rate limiting + Monitoring
+# Encryption + Audit logging
+```
+
+### 3. Monitor and Alert
+
+```ruby
+# Log security events
+# Monitor for attacks
+# Alert on violations
+# Regular security audits
+```
+
+### 4. Follow Security Standards
+
+```ruby
+# Use HTTPS
+# Implement proper access controls
+# Follow OWASP guidelines
+# Regular security updates
+```
+
+## Next Steps
+
+Now that you understand security considerations:
+
+1. **Explore [Performance](performance.md)** for security-performance trade-offs
+2. **Check out [Algorithms](algorithms.md)** for technical security details
+3. **Review [Configuration](configuration.md)** for secure configuration options
+4. **Read [API Reference](api-reference.md)** for complete security documentation