Backend: memory

The memory backend stores data in RAM, providing the fastest access but with no persistence across provider restarts.

When to use

Use the memory backend when:

• You need maximum performance

• Data is temporary or can be regenerated

• You’re implementing caching

• You’re testing or prototyping

• Persistence is not required

Characteristics

Fast: All operations are in-memory with no I/O overhead

Volatile: Data is lost when the provider stops

No size limits: Limited only by available RAM

Simple: No configuration needed

Configuration

In Bedrock configuration:

{
    "providers": [{
        "type": "warabi",
        "provider_id": 42,
        "config": {
            "target": {
                "type": "memory",
                "config": {}
            }
        }
    }]
}

The memory backend has no configuration options - just use an empty config object.

In C++ code:

#include <warabi/Provider.hpp>

auto config = R"(
{
    "target": {
        "type": "memory",
        "config": {}
    }
}
)";

warabi::Provider provider(engine, 42, config);

Memory management

The memory backend allocates memory dynamically as regions are created and written to:

// Create region (no allocation yet)
warabi::RegionID id;
target.create(&id);

// Write data (memory allocated now)
std::vector<char> data(1024 * 1024);  // 1 MB
target.write(id, 0, data.data(), data.size());

Memory is freed when: - A region is destroyed: target.erase(id) - The provider is shut down