Getting Started =============== Portability ------------ ndnSIM 2.0 has been successfully compiled and used on following platforms: - Ubuntu Linux 12.04 (see the note) - Ubuntu Linux 14.04 - OS X 10.10 .. note:: ndnSIM is currently cannot be compiled on Ubuntu Linux 12.04 with the packaged boost libraries (there is an `issue with boost 1.48 and gcc 4.6 `_). It is still possible to compile ndnSIM on this platform, but either compiler or boost libraries (or both) need to get upgraded. More recent version of boost can be installed from "Boost C++ Libraries" team PPA:: sudo apt-get install python-software-properties sudo add-apt-repository ppa:boost-latest/ppa sudo apt-get update sudo apt-get install libboost1.55-all-dev # add --boost-libs=/usr/lib/x86_64-linux-gnu to ./waf configure for ndn-cxx and ns3 # ./waf configure --boost-libs=/usr/lib/x86_64-linux-gnu Make sure that all other version of boost libraries (``-dev`` packages) are removed, otherwise compilation will fail. .. _requirements: Prerequisites ------------- 1. `ndnSIM also required ndn-cxx library and all of its prerequisites `_. .. role:: red .. note:: :red:`ndnSIM requires boost version at least 1.49.` Many linux distribution (Fedora 16, 17 at the time of this writing) ship an old version of boost, making it impossible to compile ndnSIM out-of-the-box. Please install the latest version, following :ref:`these simple instructions `. .. note:: If you do not have root permissions to install boost, you can install it in your home folder. However, you need to be make sure that `libboost_iostreams` library is successfully compiled and is installed. Please refer to :ref:`the following example ` for the hints how to successfully compile and install boost libraries on Ubuntu Linux. 2. If you are planning to use other modules, like visualizer, a number of additional dependencies should be installed. For example, in order to run `visualizer`_ module, the following should be installed: * For Ubuntu: .. code-block:: bash sudo apt-get install python-dev python-pygraphviz python-kiwi sudo apt-get install python-pygoocanvas python-gnome2 sudo apt-get install python-rsvg ipython * For Fedora: .. code-block:: bash sudo yum install pygoocanvas python-kiwi graphviz-python # easy_install method, since pygraphviz is not (yet?) packaged into Fedora (https://bugzilla.redhat.com/show_bug.cgi?id=740687) sudo yum install graphviz-devel sudo yum install python-pip sudo easy_install pygraphviz * For OS X with MacPorts: .. code-block:: bash sudo port install py27-pygraphviz py27-goocanvas # If you add NDN macports repository, as described in # http://named-data.net/doc/NFD/current/INSTALL.html#install-nfd-using-the-ndn-macports-repository-on-os-x # you will be able to install another useful python module # sudo port install py27-kiwi * For OS X with HomeBrew .. code-block:: bash brew install boost cryptopp pkg-config libxml2 brew link --force libxml2 .. _visualizer: http://www.nsnam.org/wiki/index.php/PyViz Downloading ndnSIM source ------------------------- ndnSIM package consists of three pieces: - `ndn-cxx library `_ - a custom branch of NS-3 that contains a few useful patches - a customized python binding generation library (necessary if you want to use NS-3's python bindings and/or visualizer module) - the source code of ndnSIM module The following commands download all pieces from GitHub repositories: .. code-block:: bash mkdir ndnSIM cd ndnSIM git clone https://github.com/named-data/ndn-cxx.git ndn-cxx git clone https://github.com/cawka/ns-3-dev-ndnSIM.git ns-3 git clone https://github.com/cawka/pybindgen.git pybindgen git clone https://github.com/named-data/ndnSIM.git ns-3/src/ndnSIM The few modification to the base NS-3 code are necessary to run ndnSIM, and the code is periodically synchronized with the official developer branch. Eventually, all the changes will be merged to the official branch, but for the time being, it is necessary to use the customized branch. Compiling and running ndnSIM ---------------------------- - Compile and install ndn-cxx library .. code-block:: bash cd ndnSIM/ndn-cxx ./waf configure ./waf sudo ./waf install .. note:: On Ubuntu platform you can also install ndn-cxx library from `NDN PPA repository `_ .. code-block:: bash sudo apt-get install ndn-cxx-dev If you are using ndn-cxx from PPA, please make sure that you do not have the existing installation of ndn-cxx library compiled from source (e.g., in /usr/local). Otherwise, the version mistmach may result in compilation errors. - Compile NS-3 with ndnSIM module ndnSIM uses standard NS-3 compilation procedure. Normally the following commands should be sufficient to configure and build ndnSIM with python bindings enabled: .. code-block:: bash cd ./waf configure --enable-examples ./waf On MacOS (with macports), you may need to modify the configure command to use macports version of python: .. code-block:: bash cd ./waf configure --with-python=/opt/local/bin/python2.7 --enable-examples # or run ``sudo port select python python27`` ./waf .. note:: On OS X configuration stage may get :ref:`stuck at detecting gtk module `. Make sure you have `XQuartz `_ installed or disable python as described in the following instructions. Python bindings is an optional and not very stable feature of NS-3 simulator. It is possible to disable python bindings compilation either to speed up compilation or to avoid certain compilation errors (e.g., "Could not find a task generator for the name 'ns3-visualizer'"): .. code-block:: bash cd ./waf configure --disable-python --enable-examples ./waf For more configuration options, please refer to ``./waf --help``. Simulating using ndnSIM ----------------------- - Examples simulations When NS-3 is compiled with ``--with-examples`` flag, you can directly run all examples described in :doc:`examples section of this tutorial `. For example, to run ``ndn-simple.cpp`` scenario, you can run the following command: .. code-block:: bash ./waf --run=ndn-simple To run ``ndn-grid.cpp`` scenario: .. code-block:: bash ./waf --run=ndn-grid To run the sample simulation scenarios with the logging module of NS-3 enabled (note that this will work only when NS-3 is compiled in debug mode): .. code-block:: bash NS_LOG=ndn.Producer:ndn.Consumer ./waf --run= If you have compiled with python bindings, then you can try to run these simulations with visualizer: .. code-block:: bash ./waf --run=ndn-simple --vis or: .. code-block:: bash ./waf --run=ndn-grid --vis .. note:: Do not forget to configure and compile NS-3 in optimized mode (``./waf configure -d optimized``) in order to run actual simulations. - Real experimentation While it is possible to write simulations directly inside NS-3 (in ``scratch/`` folder) or ndnSIM (in ``examples/``), the recommended way is to write your simulation scenarios, as well as any custom extensions, separately from the NS-3 or ndnSIM core. For example, you can use the following template to write your extensions, simulation scenarios, and metric processing scripts: ``_: .. code-block:: bash mkdir ndnSIM cd ndnSIM git clone git://github.com/cawka/ns-3-dev-ndnSIM.git ns-3 git clone git://github.com/cawka/pybindgen.git pybindgen git clone git://github.com/NDN-Routing/ndnSIM.git ns-3/src/ndnSIM # Build and install NS-3 and ndnSIM cd ns-3 ./waf configure -d optimized ./waf sudo ./waf install cd .. git clone git://github.com/cawka/ndnSIM-scenario-template.git scenario cd scenario export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH ./waf configure ./waf --run For more detailed information, refer to `README file `_. Examples of template-based simulations ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1. ndnSIM examples from ``_ website and more: - ``_, or 2. Script scenarios and graph processing scripts for simulations used in "A Case for Stateful Forwarding Plane" paper by Yi et al. (``_): - ``_, or 3. Script scenarios and graph processing scripts for simulations used in "Rapid Traffic Information Dissemination Using Named Data" paper by Wang et al. (``_): - ``_, or - Rocketfuel-based topology generator for ndnSIM preferred format (randomly assigned link delays and bandwidth, based on estimated types of connections between nodes): - ``_, or