Overview

In-Memory Operation mode enables LiteGraph to run entirely in system memory, providing exceptional performance and ultra-low latency for high-throughput scenarios. This mode is ideal for applications requiring maximum speed, real-time processing, and temporary data storage with optional persistence capabilities.

Key benefits include:

Ultra-High Performance: Maximum speed with data stored in RAM
Low Latency: Minimal response times for all operations
Real-Time Processing: Ideal for streaming and real-time applications
Temporary Storage: Perfect for caching and session management
Optional Persistence: Manual control over data persistence to disk
Memory Management: Efficient memory usage and garbage collection
Scalability: Enhanced performance for high-concurrency scenarios

Configuration

For high-performance scenarios, run the database entirely in memory by configuring the following settings:

Enable In-Memory Mode

Configure In-Memory Operation: Set InMemory to true in your litegraph.json configuration file:

{
  "LiteGraph": {
    "InMemory": true,
  }
}

Manual Flush to Disk

Important: When running in-memory mode, data is stored only in RAM by default. You must manually flush data to disk to ensure persistence:

curl -X POST -H "Authorization: Bearer litegraphadmin" \
     -H "Content-Type: application/json" \
     http://localhost:8701/v1.0/flush

import { LiteGraphSdk } from "litegraphdb";

var api = new LiteGraphSdk(
  "http://localhost:8701/",
  "<Tenant-Guid>",
  "*******"
);

const flushToDisk = async () => {
  try {
    const data = await api.Admin.flush();
    console.log(data, "Data flushed to disk successfully");
  } catch (err) {
    console.log("Error flushing data:", JSON.stringify(err));
  }
};

import litegraph

sdk = litegraph.configure(
    endpoint="http://localhost:8701",
    tenant_guid="Tenant-Guid",
    access_key="******",
)

def flush_db_to_disk():
    backup = litegraph.Admin.flush_db_to_disk()
    print(backup)

flush_db_to_disk()

Performance Considerations

Memory Usage

RAM Requirements: Ensure sufficient system memory for your dataset
Memory Monitoring: Monitor memory usage to prevent system overload
Data Size Limits: Consider dataset size relative to available RAM
Concurrent Operations: In-memory mode handles high concurrency efficiently

Use Cases

Ideal for:

Real-time analytics and reporting
High-frequency trading systems
Session management and caching
Temporary data processing
Development and testing environments
High-throughput API services

Not recommended for:

Large datasets exceeding available RAM
Long-term data storage without persistence
Systems with limited memory resources
Production environments requiring automatic persistence

Best Practices

When using in-memory operation mode, consider the following recommendations:

Memory Monitoring: Implement monitoring to track memory usage and prevent overflow
Regular Flushing: Establish a schedule for regular data flushing to disk
Backup Strategy: Create backups before critical operations or system restarts
Error Handling: Implement proper error handling for memory-related issues
Resource Planning: Plan memory allocation based on expected data volumes
Testing: Thoroughly test in-memory operations in your specific environment
Documentation: Document flush procedures and memory management strategies

Next Steps

After configuring in-memory operation, you can:

Implement automated flush scheduling and management
Set up memory monitoring and alerting systems
Create data persistence strategies and procedures
Optimize memory usage and garbage collection settings
Develop backup and recovery procedures for in-memory data
Implement performance monitoring and optimization
Create comprehensive memory management documentation