Local Platform

Private Cloud

Introduction

The Private Cloud add-on to the Local Platform enables a cluster of centrally managed servers to run work tasks from a distributed team of users. This is much more efficient than each quant having a full copy of data or their own GPU resources on their individual computers.

Key Features

  • Share access to centralized servers with expensive specialized hardware. Set up a powerful server with plenty of RAM, CPU, and GPU resources for your team to build and train the latest models.
  • Provide distributed teams with safe, monitored access to proprietary data. Centralizing data allows your team to access it without the potential for distribution.
  • Keep infrastructure costs low by reducing total hardware expenses and making better use of hardware.

Requirements

To use a Private Cloud, you need an Institutional Subscription to the Local Platform with the Private-Cloud Add-On. This is licensed per server and supports scaling up to 10 servers. If you're interested in this solution, make an appointment with our team to discuss if it will work for your fund.

For full pricing information, see the Pricing page.

Infrastructure Layout

Th following image describes the Private cloud infrastructure:

Clients run research notebooks, backtests, optimizations, or live trading strategies on their laptops using the QuantConnect extension in VSCode. The work runs on the centralized servers utilizing the centralized data and powerful servers inside the corporate local network. Results are streamed back to the Local Platform user interface so the quant can research as before.

Deploy a Private Cloud

To deploy a private cloud, follow these steps:

Set up a Master Work Server

  1. Install and authenticate the LEAN CLI.
  2. Run the lean deploy private-cloud --master LEAN CLI command to start a work queue and listen for jobs from the team.
  3. If you are setting up a single server, this is all you need to do.

  4. To expand the private cluster further so work is distributed across many servers, install the LEAN CLI into the second server and run lean deploy private-cloud --slave.
  5. This will spin up the second work-server to consume jobs.

Set up a Team Thin-Client

  1. On the quant analyst's computer, install the Local Platform.
  2. Install and authenticate the LEAN CLI.
  3. Open a QuantConnect project inside the VSCode extension and select the Private Cloud deployment target.
  4. Enter the IP/DNS of the master-work server provided when running the LEAN CLI start-up command.

Presale

This technology has not yet been released and is currently open for discounted presales. If you are interested in using this with your organization, book an appointment to discuss how it could scale an on-premise quant infrastructure.

You can also see our Videos. You can also get in touch with us via Discord.

Did you find this page helpful?

Contribute to the documentation: