System Requirements

This document describes the system requirements and recommended configurations for running Scintirete.

📋 Basic Requirements

Operating System Support

Scintirete supports the following operating systems:

Hardware Requirements

Minimum Configuration

• CPU: 1 core (x86_64 or ARM64)
• Memory: 512 MB RAM
• Storage: 100 MB available disk space
• Network: TCP/IP support

Recommended Configuration

• CPU: 2+ cores (x86_64 or ARM64)
• Memory: 2 GB+ RAM
• Storage: 1 GB+ SSD storage
• Network: Gigabit Ethernet

🧮 Memory Calculation Guide

Vector Storage Memory Requirements

Scintirete stores all vector data in memory to ensure optimal performance. Memory requirements are primarily determined by the following factors:

Total Memory Required = Vector Data + HNSW Index + Metadata + System Overhead

Vector Data Memory

Vector Data Memory = Number of Vectors × Vector Dimension × 4 bytes (float32)

Example Calculations:

• 1 million 768-dimensional vectors: 1,000,000 × 768 × 4 = ~2.9 GB
• 1 million 1536-dimensional vectors: 1,000,000 × 1536 × 4 = ~5.7 GB

HNSW Index Memory

HNSW index typically requires an additional 30-50% memory overhead:

Index Memory ≈ Vector Data Memory × 0.3 to 0.5

Complete Memory Estimation Table

Memory Optimization Recommendations

1. Reserve Memory: Reserve at least 25% of memory for system and other processes
2. Batch Import: Import large amounts of data in batches to avoid memory peaks
3. Monitor Usage: Use the /metrics endpoint to monitor memory usage
4. Regular Rebuild: Periodically rebuild indexes to clean up marked-deleted vectors

🔧 Runtime Dependencies

No External Dependencies

Scintirete is a statically compiled single binary with no additional runtime dependencies required:

• ✅ No Python Environment Required
• ✅ No Go Runtime Required
• ✅ No Database Software Required
• ✅ No Complex System Libraries Required

Optional Dependencies (for Integration Features)

🌐 Network Requirements

Port Configuration

Firewall Configuration

Minimal configuration (local access only):

# Allow local loopback access
sudo ufw allow from 127.0.0.1

Internal network service configuration:

# Allow internal network access to HTTP API
sudo ufw allow from 192.168.0.0/16 to any port 8080

# Allow internal network access to gRPC
sudo ufw allow from 192.168.0.0/16 to any port 9090

Public network service configuration (not recommended, use reverse proxy and TLS):

# Caution: Only open public access with appropriate security measures
sudo ufw allow 8080
sudo ufw allow 9090

External Network Requirements

If using text vectorization features, access to OpenAI API is required. Please configure base_url and api_key accordingly.

💾 Storage Requirements

Disk Space Planning

Total Storage Required = Data Files + Log Files + Configuration Files + System Overhead

Data Storage Estimation

Storage Configuration Recommendations

File System Recommendations

🔒 Security Requirements

Network Security

1. Firewall: Configure appropriate firewall rules
2. TLS: Use HTTPS/TLS in production environments
3. Authentication: Configure strong passwords or key authentication
4. Network Isolation: Place database services in protected network segments

File System Permissions

# Create dedicated user (recommended)
sudo useradd -r -s /bin/false scintirete

# Set data directory permissions
sudo mkdir -p /var/lib/scintirete
sudo chown scintirete:scintirete /var/lib/scintirete
sudo chmod 750 /var/lib/scintirete

# Set configuration file permissions
sudo chmod 640 /etc/scintirete/scintirete.toml
sudo chown root:scintirete /etc/scintirete/scintirete.toml

🚀 Performance Tuning

System-Level Optimization

Linux System Parameters

# Increase file descriptor limits
echo "* soft nofile 65536" >> /etc/security/limits.conf
echo "* hard nofile 65536" >> /etc/security/limits.conf

# Optimize memory management
echo 'vm.swappiness = 1' >> /etc/sysctl.conf
echo 'vm.dirty_ratio = 15' >> /etc/sysctl.conf
echo 'vm.dirty_background_ratio = 5' >> /etc/sysctl.conf

# Apply configuration
sysctl -p

Docker Optimization

# Limit container memory (prevent OOM)
docker run -m 4g scintirete/scintirete:latest

# Use local SSD volume
docker run -v /fast-ssd/scintirete:/app/data scintirete/scintirete:latest

Application-Level Optimization

1. HNSW Parameter Tuning: Refer to HNSW Parameter Tuning
2. Batch Operations: Use batch insert and search interfaces
3. Connection Pooling: Client-side connection pool management
4. Monitoring-Based Tuning: Adjust configuration based on Prometheus metrics

📊 Environment Verification

System Information Check

# Check basic system information
uname -a
cat /etc/os-release

# Check hardware resources
free -h              # Memory
df -h               # Disk space
lscpu               # CPU information

Network Connectivity Test

# Test port availability
netstat -tuln | grep 8080
netstat -tuln | grep 9090

# Test network latency
ping your-scintirete-server

🔧 Deployment Mode Comparison

Choose the appropriate deployment mode to meet your requirements and resource constraints.