Backend.AI

Backend.AI is a streamlined backend API server that hosts heterogeneous programming languages and popular AI frameworks. It manages the underlying computing resources for multi-tenant computation sessions where such sessions are spawned and executed instantly on demand.

Server-side Components

If you want to run a Backend.AI cluster on your own, you need to install and configure the following server-side components. All server-side components are licensed under LGPLv3 to promote non-proprietary open innovation in the open-source community.

There is no obligation to open your service/system codes if you just run the server-side components as-is (e.g., just run as daemons or import the components without modification in your codes). Please contact us (contact-at-lablup-com) for commercial consulting and more licensing details/options about individual use-cases.

For details about server installation and configuration, please visit our documentation.

Manager with API Gateway

It routes external API requests from front-end services to individual agents. It also monitors and scales the cluster of multiple agents (a few tens to hundreds).

• Package namespace: ai.backend.gateway and ai.backend.manager
• https://github.com/lablup/backend.ai-manager

Agent

It manages individual server instances and launches/destroys Docker containers where REPL daemons (kernels) run. Each agent on a new EC2 instance self-registers itself to the instance registry via heartbeats.

• https://github.com/lablup/backend.ai-agent
  • Package namespace: ai.backend.agent
• https://github.com/lablup/backend.ai-accelerator-cuda (CUDA accelerator plugin)
  • Package namespace: ai.backend.acceelrator.cuda

Server-side common plugins (for both manager and agents)

• https://github.com/lablup/backend.ai-stats-monitor
  • Statistics collector based on the Datadog API
  • Package namespace: ai.backend.monitor.stats
• https://github.com/lablup/backend.ai-error-monitor
  • Exception collector based on the Sentry API
  • Package namespace: ai.backend.monitor.error

Kernels

A set of small ZeroMQ-based REPL daemons in various programming languages and configurations.

• https://github.com/lablup/backend.ai-kernel-runner
  • Package namespace: ai.backend.kernel
  • A common interface for the agent to deal with various language runtimes
• https://github.com/lablup/backend.ai-kernels
  • Runtime-specific recipes to build the Docker images (Dockerfile)

Jail

A programmable sandbox implemented using ptrace-based sytem call filtering written in Go.

• https://github.com/lablup/backend.ai-jail

Hook

A set of libc overrides for resource control and web-based interactive stdin (paired with agents).

• https://github.com/lablup/backend.ai-hook

Commons

A collection of utility modules commonly shared throughout Backend.AI projects.

• Package namespaces: ai.backend.common
• https://github.com/lablup/backend.ai-common

Client-side Components

Client SDK Libraries

We offer client SDKs in popular programming languages. These SDKs are freely available with MIT License to ease integration with both commercial and non-commercial software products and services.

• Python (provides the command-line interface)
  • pip install backend.ai-client
  • https://github.com/lablup/backend.ai-client-py
• Java
  • Currently only available via GitHub releases
  • https://github.com/lablup/backend.ai-client-java
• Javascript
  • npm install backend.ai-client
  • https://github.com/lablup/backend.ai-client-js
• PHP (under preparation)
  • composer require lablup/backend.ai-client
  • https://github.com/lablup/backend.ai-client-php

Media

The front-end support libraries to handle multi-media outputs (e.g., SVG plots, animated vector graphics)

• The Python package (lablup) is installed inside kernel containers.
• To interpret and display media generated by the Python package, you need to load the Javascript part in the front-end.
• https://github.com/lablup/backend.ai-media

Integrations with IDEs and Editors

• Jupyter Kernel: data scientists' favorite tool
  • Package namespaces: backend.ai-integration-jupyter
  • https://github.com/lablup/backend.ai-integration-jupyter
• Visual Studio Code Extension
  • Search “Live Code Runner” among VSCode extensions.
  • https://github.com/lablup/vscode-live-code-runner
• Atom Editor plugin
  • Search “Live Code Runner” among Atom plugins.
  • https://github.com/lablup/atom-live-code-runner
• IntelliJ / PyCharm plugin
  • Coming soon!