Link Search Menu Expand Document

Install the MDI

Table of Contents

  1. Stage 1 Pipelines vs. Stage 2 Apps
  2. Local, Remote, and Public Apps Servers
    1. Local mode
    2. Remote modes
    3. Public server mode

This section describes the paths to installing the MDI itself, i.e., the frameworks. It is easy, but there are several ways to do it. The following are important decision points.

Stage 1 Pipelines vs. Stage 2 Apps

Review the MDI’s analysis stages to decide what kind of tools you seek to run.

If you will run Stage 1 Pipelines from the command line of an HPC server, install and use our ‘mdi’ utility:

If you will run Stage 2 Apps on data generated elsewhere, including the Pipeline Runner app, do that on your personal computer using the MDI Desktop (recommended):

or directly within an R console (not recommended for most users):

or set up a public web server on Amazon Web Services (AWS) using our Amazon Machine Images (AMIs):

Local, Remote, and Public Apps Servers

Stage 1 Pipelines nearly always run on an HPC server as linked above. However, there are many ways to run a Stage 2 Apps server depending on your needs.

Local mode

In local mode, you run the R Shiny server on your desktop or laptop computer, which acts as both the web server and client (i.e., browser).

First, you must install R:

Then, the best path is to use the MDI Desktop to install and run the server (recommended):

Alternatively, install the MDI from within R as described here (not recommended for most users):

The advantages of this mode are security and speed.

Remote modes

You can also use the MDI Desktop to install and run a MDI server on a remote HPC server, or one of its nodes, with a connection over a secure, SSH port tunnel.

An advantage of this method is that your apps can have ready access to the same disk drives as your pipelines.

Public server mode

As noted above, advanced users can also set up a public Stage 2 web server using our AWS AMIs:

as described in the following repositories:

Here, the main advantage is stable, shared access between many users.