First, set the following:


Then, choose below:

Quick Start Guides

Example Tour

Reference Manual

AMPL-Gurobi Guide

Remote Services

Cloud Guide

Gurobi Instant Cloud

Gurobi Instant Cloud

Gurobi Instant Cloud allows you to start and stop Gurobi Compute Servers on the cloud. You can start multiple machines without the need for your own hardware or local Gurobi licenses. Computations are seamlessly offloaded to these servers. Depending on your cloud license type, these machines provide the full set of Compute Server features, including queuing, load balancing, and distributed parallel computing.


When using the Instant Cloud, there are always three systems involved: your client machine, the Instant Cloud Manager, and a cloud Compute Server.

The program that requests a Gurobi Instant Cloud machine and submits optimization models to this server runs on your client machine. Note, however, that this program does not actually need to be aware that it will be using Gurobi Instant Cloud. You have a few options for configuring the client to use the Instant Cloud. The simplest and most seamless is to set up a cloud license file. The alternative is to use a programming language API, which gives your program additional control over how it uses the cloud. Details on launching cloud machines from your client program follow shortly.

The Instant Cloud Manager manages the configuration and launching of cloud machines. Your client program will send credential information to this website, along with a request to launch an Instant Cloud machine. The specific action taken in response to this request depend on configuration information that you manage through the website. For each license, you set up things like the number of servers to launch, the types and geographic regions of these machines, the maximum number of simultaneous jobs to run on each server, etc.

Once the Instant Cloud Manager launches the requested Compute Servers, it passes information about these servers back to your client program. The client program then directly interacts with the servers, sending the Gurobi model, launching a solve on the model, requesting solution information, etc. As with any Gurobi Compute Server, this process is entirely transparent to the client program.

Now that we've given a high-level description of the overall process, we need to cover a few important details.