APE Logo

Ad hoc Protocol Evaluation testbed

This is the homepage of the APE testbed project. APE testbed is short for Ad hoc Protocol Evaluation testbed. The APE project started as a research project at the Department of Computer Systems at Uppsala University and was partly funded by Ericsson. APE is now released under GNU General Public License . Its intended use is for evaluation of mobile ad hoc routing protocols. APE aims at making the process of performing complex real-world tests as easy as possible. It focuses on smooth deployment, high ability of customization and ability to easily run several routing protocol implementations for comparisons. Furthermore, APE provides means to assess the similarity between repeated testruns.

For detailed information and research papers/publications, see section APE documentation .

APE project members: Erik Nordström, Henrik Lundgren, David Lundberg, Christian Tschudin, Per Gunningberg.

SourceForge Logo


What exactly is APE?

What can APE do for me?


APE tools




Related links


What exactly is APE?

APE is our contribution to the research community in the area of ad hoc networking. It is free to use and modify by anyone as long as it is in line with the GPL.

There is no clear definition of what a testbed is or what it comprises. APE however, can be seen as containing two things:

The APE Linux distribution is the environment where tests are performed and data is gathered. It is self-contained in a small package with a size of about 8 Mb, and can be installed in an existing Linux or Windows environment (with DOS support), without re-partitioning or complex installation procedures. Basically just download, install, and run. This simple installation has allowed us to use students and their laptops as participants in our tests, and we have been able to scale tests to include more than 30 nodes.

The APE distribution comes pre-configured and contains tools for data gathering at both IP and ethernet level. It also has a scenario scripting system, which enables APE testruns to use movement choreography, with instructions to participants. Scenarios also define traffic generated during tests. This allows testruns to be repeated in a very precise manner, so that results can be verified.

The APE source code allows you to build customized APE distribution packages, with new scenarios, protocols and much more. The source package also contain analysis tools for analysis of gathered data.

APE is currently prepared to support the following protocol implementations:
This list does not contain the only protocols which at this time can run in APE.  APE is a Linux environment, therefore any Linux based protocol implementation should run in APE, granted it is built with the right kernel version or other software necessary. A supported protocol is easy to compile and include when building a customized APE distribution package and are also available in our scenario system. Adding support for a new protocol is easy, so expect more extensive protocol support as implementations become available.

What can APE do for me?

If you are a teacher running a project about ad-hoc networking in your computer science course, then APE can be a good way for students to do experiments with ad-hoc network protocols.

If you are conducting real-world research in the ad-hoc networking area, then APE may be a good way to get the most out of your efforts. It can simplify time consuming tasks, like installation and configuration of execution environments, and it provides good tools for data gathering and analysis to build upon.


APE Tools


Mackill is a MAC filter integrated into the APE testbed. With Mackill you can force different connectivity configurations in the ad-hoc network - without nodes required to be physically separated. This greatly simplifies testing of, for example, routing algorithms.

In APE, Mackill is also scriptable in our scenario system. This makes it possible to create scenarios that plays out automatically. Mackill is a very powerful tool for doing side by side, repeatable test of different routing algorithms. However, for obvious reasons, a Mac filter is not a substitute for tests with real movements.


APE includes tools written in Perl for simple analysis of:

There is also an (experimental) graphical frontend available.

Analysis tool


APE-view is a log driven animation tool included in the APE source package. It displays the topological configuration of nodes during testruns, based on logged signal strength between nodes. Below are some screenshots from a scenario we call "Double Split" with 34 nodes. Yellow dots are nodes, and lines show the connectivity between them.


APE-view is a good way to verify the topological configuration aimed for in a scenario.


If you find APE useful and you start to modify APE for your own purposes, we urge you to contribute your work back to the APE project. This will allow APE to evolve and others to benefit from your work.
Example contributions could be:

APE documentation

A technical report about the APE testbed. A version of this paper is accepted for publication at WCNC'02.
Documentation covering how to build, install and run the APE testbed is available here [.ps] [ .pdf ].
This document includes the hardware/software requirements for running the APE testbed.


APE binary distribution

The latest binary distribution package is available here. This is a compiled APE package which can be used to try APE and get a feeling for what it is. Remember though, that this is only the execution environment used when performing testruns. For analysis tools, build and customization scripts you really need the source package.

APE source code

The latest source is available here

Source contents:

Related links

AODV-UU - Our own AODV implementation

LUNAR - Lightweight Underlay Network Ad-hoc Routing

ARRCANE - Active Routing and Resource Control in Ad-hoc NEtworks

Mad-hoc - AODV implementation

OLSR implementation from INRIA

APE contact

Mailing lists at the APE testbed project page at SourceForge.


Page last modified 2002-01-10