OpenDsec: self-hosted sandboxes for RL rollouts and code-running agents

Описание и применение

OpenDsec — это самодостаточная платформа для изолированных сред выполнения, позволяющая быстро запускать микровМ, форкать их с памятью и приостанавливать задачи, что упрощает обучение агентных моделей с усилением.

Задача

Agentic reinforcement learning training requires thousands of isolated code execution environments that can be paused, resumed, and forked with memory, but hosted sandbox services are limited and costly.

Результат

OpenDsec delivers microVMs that boot in ~0.1 s, fork with memory in ~10 ms, and support task-level pause, enabling efficient RL rollouts on bare-metal or cloud VMs.

Рабочий процесс

• Run the install script to build and deploy OpenDsec components
• Import a base image (e.g., python:3.12) from Docker Hub
• Create a sandbox via the SDK (DSecClient.run_sandbox)
• Execute code inside the sandbox with run_shell
• Fork the sandbox with memory using fork(count=N)
• Stop all sandboxes with client.stop()

Доказательства

If you train a model with agentic reinforcement learning, every rollout needs somewhere to run code: install a package, run the tests, hit the network, write files.

Источник

The program is the whole client side: ... print(6*7) ...

Источник

Fast microVMs. About 0.1 s for the VM to boot, about 55 ms to create one from a memory snapshot, about 50 ms to resume one from a pause.

Источник

One 384-thread node took a burst of 5120 microVMs in 37 s. A two-node cluster took 8192 containers in one burst with no failures.

Источник