Version 25 (modified by, 9 years ago) (diff)


GENI OpenFlow Experiment Example 2

This page illustrates an OpenFlow (OF) example for the Omni How-to page. OpenFlow is used for the control of the network resources in the GENI backbone and at meso-scale campus locations. Experimenters can use these GENI resources without learning OpenFlow. The GPO can help arrange connectivity across your GENI slice, including IP connectivity and/or layer 2 VLANs.

For experimenters who desire control over the network switching fabric in their slices, OpenFlow provides increased flexibility. Your experiment will incorporate an OpenFlow controller. Many experimenters use existing controller code as-is, some use controllers with relatively simple modifications, and some experimenters have written their own controller. The GPO is available to help design your controller and plan your experiment configuration. Feel free to contact us at

This experiment uses credentials from the GPO ProtoGENI Clearinghouse ( to set up a 3 Meso-scale sites OpenFlow experiment. The sites are includes BBN, Wisconsin and Clemson each with a MyPLC compute resource. If you are interested in using PG resources, see the HelloGENI experiment page.

No image "OF-Experiment-example.jpg" attached to GENIExperimenter/ExperimentExample-OF-FOAM

Omni tools are used to reserve resources in each of the site's FlowVisor OpenFlow Aggregate Manager (FOAM) and GENI Aggregate Managers. Steps required to set up the 3 Sites OpenFlow experiment are detailed in this page:

  • Authentication and Omni tools configuration
  • Determine Resources Availability
  • Define Compute Resources RSpecs
  • Define OpenFlow Network Resources Rspecs
  • Create Slice and Request Resources
  • Run Experiment
  • Release Resources

Note: FOAM sites must be federated in order to trust a Clearinghouse. This is an OF administrator task that must take place before any of the user OF operations can take place.

Authentication and Omni tools configuration

This example assumes that GENI credentials have been set up as detailed in the GENI Credentials section of the Omni how-to page.

The following omni_config was used for this experiment along with the credentials from the GPO Clearinghouse to access the GENI Aggregate Managers in this experiment:

default_cf = pgeni
users = lnevers

# ---------- Users ----------
urn =
keys = ~/.ssh/

# ---------- Frameworks ----------
type = pg
ch =
sa =
cert = ~/.ssl/pgeni/encrypted-cleartext.pem
key = ~/.ssl/pgeni/encrypted-cleartext.pem

ch =
sa =
cert = ~/.ssl/protogeni/encrypted-cleartext.pem
key = ~/.ssl/protogeni/encrypted-cleartext.pem

A list of Aggregate managers is compiled in the GENI Aggregate page, if your site is not listed on this page you should contact your site administrator, or

Reserve GENI IP subnet

In the Meso-scale network a layer 3 experiment requires its own separate IP subnet in order to control all traffic without interfering with traffic from other experiments. At this point you need to request an IP subnet from the mail list. The request should also specify a duration.

Determine Resources Availability

Each of the 3 sites will requires an OpenFlow RSpec and a compute resource RSpec. Also an OpenFlow RSpec will be required for the GENI Core (NLR in this example). To find out what resources are available use Omni tools to get a list of available resources. Following is a list of MyPLC and FOAM aggregates available in the meso-scale network:

Site MyPLC Aggregate URL FOAM Aggregate URL
GA Tech

For complete list of available aggregates see the available GENI Aggregates page.

Define Compute Resources RSpecs

This experiment uses Omni V1.6.1 to set up the experiment. As a first step, the experimenter determines the available resources at each of the MyPLC Aggregate Managers. Following are the omni commands issued to get the manifests for available compute resources:

  1. Clemson MyPLC aggregate:
     ./src/ -a listresources -o            # generates rspec-myplc-clemson-edu-12346-.xml
  2. Wisconsin MyPLC aggregate:
     ./src/ -a  listresources -o # rspec-wings-openflow-1-wail-wisc-edu-12346-.xml
  3. BBN MyPLC aggregates:
     ./src/ -a listresources -o            # generates rspec-myplc-gpolab-bbn-com-12346-.xml

Each of the omni command above used the -o argument to get the listresources output to a file, all output files are attached to this page as a reference. The Advertisement RSpec generated by the listresources are then reviewed to generate the request RSpecs to reserve the resources. Following are the compute resources RSpec defined based on the output of the listresources commands above.

BBN MyPLC Compute Resource RSpec

<?xml version="1.0"?>
This rspec will create a container on each of the BBN production MyPLC plnodes.

<RSpec type="SFA">
  <network name="plc.gpolab">
       <node component_id="" 
       node_id="n1" site_id="urn:publicid:IDN+plc:gpolab:gpolab+authority+sa">      
            <hostname>      </hostname>      

Clemson MyPLC Compute Resources RSpec

<?xml version="1.0" ?>
<!-- Resources at AM:
        URN: unspecified_AM_URN
<RSpec type="SFA">  
    <network name="plc">    
        <node component_id="" 
                node_id="n1" site_id="urn:publicid:IDN+plc:cu+authority+sa">    
            <hostname>      </hostname>      

Wisconsin MyPLC Compute Resources RSpec

<?xml version="1.0" ?>
<!-- Resources at AM:
        URN: unspecified_AM_URN
<RSpec type="SFA">  
    <network name="plc">    
        <node component_id="" 
                node_id="n9" site_id="urn:publicid:IDN+plc:wisc+authority+sa">  
            <hostname>      </hostname>   

Define OpenFlow Network Resources Rspecs

This section defines how to figure out the OF network path for youre rspec. To write an OF RSpec you must determine the end-to-end OF path that will be used for the experiment. Following is an outline of the process to determine the content of your OF RSpec:

Step 1. Collect listresources from all FOAM Aggregate. The manifest RSpec details all resources at the aggregate, but does not provide any insight into whether they are available. Step 2. Review GeniAggregate to determine how each MyPLC host maps to the Site's OpenFlow switch.

ports are

Step 3. Using guidelines from the Writing OF V3 RSpec page merge the information from step 1 and 2 to generate the request RSpecs.

Each of the above is covered in detail below:

Step 1 . Collect listresources from all FOAM Aggregate:

  1. Clemson FOAM and MyPLC aggregates:
     ./src/ -a listresources -o   # generates output file rspec-foam-clemson-edu-3626-foam-gapi-1.xml
  2. Wisconsin FOAM and MyPLC aggregates:
     ./src/ -a listresources -o    # generates rspec-foam-wail-wisc-edu-3626-foam-gapi-1.xml
  3. BBN FOAM and MyPLC aggregates:
     ./src/ -a listresources -o         # generates rspec-foam-gpolab-bbn-com-3626-foam-gapi-1.xml
  4. NLR OpenFlow Core FOAM server:
    $ ./src/ -a -o listresources           # generates rspec-foam-nlr-net-3626-foam-gapi-1.xml

Information in these files provides the switch and the port information to be used in making the OF flowspace reservation.

Step 2 . Review GeniAggregate to determine MyPLC hosts mapping to the Site's OpenFlow switch.

Step 3 . Merge Step 1 and 2 to generate request RSpecs.

Determine OF network path

From the listresources output files (attached to this page) determined the sites DPIDs BBN FOAM - Aggregate URL FOAM Clemson - Aggregate FOAM NLR - Aggregate URL FOAM Wisconsin - Aggregate URL

OpenFlow Rspecs for FOAM

BBN OpenFlow Network Resources RSpec

Clemson OpenFlow Network Resources RSpec

Wisconsin OpenFlow Network Resources RSpec

NLR OpenFlow Network Resources RSpec

Create Slice and Request Resources

Using the RSpecs that define the required OF flowspace, you can now request OF resources with Omni. First create a slice using your GPO Clearinghouse credentials as defined in the omni_config. Note that the slice name should be meaningful to you and then create a sliver for each resource aggregate. In this example there are 4 OF aggregates (one per site plus one for the GENI core) and 3 compute resource aggregates (3 MyPLC), which maps to creating 7 slivers, and the following commands are executed:

 ./src/ createslice 3sites-OF-Exp
./src/ -a createsliver 3sites-OF-Exp foam_bbn.rspec
./src/ -a createsliver 3sites-OF-Exp foam_indiana.rspec
./src/ -a createsliver 3sites-OF-Exp foam_wisc.rspec
./src/ -a createsliver 3sites-OF-Exp foam_nlr.rspec
./src/ -a createsliver 3sites-OF-Exp pg-bbn.rpsec
./src/ -a createsliver  3sites-OF-Exp   myplc-clemson.rspec
./src/ -a createsliver 3sites-OF-Exp myplc-wisc.rspec

Run Experiment

Once your flowspace has been opted in you should be able to exchange traffic over the approved flows between the hosts that have been allocated to your experiment. Simply SSH into one of the hosts and ping any of the other hosts to verify connectivity. If you have connectivity between the allocated hosts, you may start your experiment, if you do not have connectivity, contact

Remember to check the expiration time for the flowspace and for each aggregate with the Omni sliverstatus and to renewslice and renewsliver to expand the duration of the resource allocation to match your experiment needs. For an example on how to check sliverstatus and renewslice see the checking status section of the other non-OF experiment.

Release Resources

When the experiment is completed you should release the resources by deleting the sliver for each of the Resource Aggregates used in the Experiment. In the example experiment used in this page, the following commands should be issued to release resources:

./src/ -a deletesliver 3sites-OF-Exp
./src/ -a deletesliver 3sites-OF-Exp
./src/ -a deletesliver 3sites-OF-Exp
./src/ -a deletesliver 3sites-OF-Exp
./src/ -a deletesliver  3sites-OF-Exp  
./src/ -a deletesliver 3sites-OF-Exp 
./src/ -a deletesliver 3sites-OF-Exp 

Email us with any questions and feedback on this page!