This repository contains code for executing the MAP system in Emulab or the DCOMP testbed as well as a lo-fi simulator.

The documentation for the lo-fi simulator can be found at (MAP-code/README.md). This simulator is meant to be run on a single system for algorithm testing.

Run ./setup to get your git repo setup with submodules.

To build use ./gradlew assemble

You will need to create the docker images for the test applications and put that in an accessible location in your test environment.
This can be done with either docker or podman. If podman is installed, it will be used. See the scripts in [Docker/service_images] for creating the images.

The deployment scripts assume that the file is in /proj/a3/map/registry/map_service_images.tar on Emulab and in /project/map/registry/map_service_images.tar on the DCOMP testbed. These locations are specified in (stage-experiment/src/main/resources/ansible/11-copy-registry-images_dcomp.yml) and (stage-experiment/src/main/resources/ansible/11-copy-registry-images_emulab.yml).

See (stage-experiment/docs/README.md) for information on staging an experiment.

See (docs/testbed-terminology.txt) for information on translating Emulab terms to DCOMP testbed terms.

Notes on using the DCOMP testbed are at (docs/dcomp-notes.md)

See (docs/topology-considerations.md) for some tips on creating and modifying topologies to run.

A modified version of iftop is included in stage-experiment/iftop. This adds an output mode that is easy to parse by the MAP code.

See (MAP-code/scenarios) for scenarios that have been used in the past. In most of the scenario directories you will find a shell script named stage.sh. This script is used to create the files that are used to run the simulation in the testbed. A typical execution of this script is ./stage.sh --experiment large --output run01 This will create files to execute the simulation in the experiment with the name large and the project map on DCOMP and a3 on Emulab. Additional options can be specified, these can be found by executing java -jar stage-experiment/build/libs/stage-experiment-{version}-executable.jar --help.

Once the experiment finishes the resulting output can be used to create various tables and charts for analysis. These can be created by executing MAP-code/src/MAP-ChartGeneration/scripts/process-hifi.sh {directory containing result tar file}. This will create a charts directory that contains all of the tables and graphs that are used for analysis.