Skip to content

Latest commit

 

History

History
57 lines (35 loc) · 2.74 KB

INSTRUCTIONS.md

File metadata and controls

57 lines (35 loc) · 2.74 KB
Raw

Rainbow-SWIM example

This document provides instructions on how to build and run the docker container for a demonstration of Rainbow and SWIM.

SWIM - A Simulator of Web Infrastructure and Management

SWIM is a self-adaptive exemplar that simulates a web application. SWIM can be used as a target system with an external adaptation manager interacting with it through its TCP-based interface or with an adaptation manager built as a simulation module. More details can be found on the SWIM website.

This demonstration uses the SWIM simulator with Rainbow as the Self-Adaptive Platform

Building SWIM-Rainbow

In the root Rainbow directory (i.e., the one that contains deployments, libs, rainbow etc., build the docker image using the following command:

> docker build -t cmuable/rainbow-example -f deployments/rainbow-example/Dockerfile .

This is a multi-stage docker build file that will first build a Rainbow release targeted for SWIM.

Running Rainbow and SWIM

  1. Run the docker container:

    > docker run -it --rm -p 6901:6901 --hostname rainbow-example cmuable/rainbow-example

  2. Log into the docker container from a browser by connecting to http://localhost:6901 (assuming you are connecting from the Docker host). You will be prompted for a password. Use vncpassword. After logging in, the screen should look like:

    Rainbow SWIM desktop

  3. To start SWIM, double-click the SWIM logo: SWIM. This launches a GUI interface from which the candidate arrival traffic profile for SWIM can be selected.

    SWIM Traffic Selection

    Double-clicking on one of the profiles will launch SWIM with that profile, starting with three initial servers all serving optional content.

  4. To start Rainbow, double-click on the Rainbow logo: Rainbow. This will bring up the Rainbow UI, which should look like:

    Rainbow UI

  5. Rainbow should start automatically - you will see the model being updated via operations:

    Model updates

  6. Eventually, Rainbow will detect that there are too many servers for the current traffic and attempt an adaptation. You should see something like:

    Rainbow adaptation This shows that a server has been removed (the two red gauge lights indicating deactivation and removal). Note that the Stitch AM window shoes that a strategy was queued (and is currently exectuing), and the Rainbow Strategy Executor is showing details of what strategy is being executed.

  7. More details about the Rainbow UI can be found in Rainbow UI Documentation