wiki:Gec5ClusterEAgenda

Version 2 (modified by Vic Thomas, 10 years ago) (diff)

--

The ORBIT Cluster met at the GEC on Monday July 20 and Tuesday July

  1. At the July 20 meeting Max Ott and Ray Raychaudhuri gave an

overview of the ORBIT testbed and discussed the integration of WiMax base stations into the cluster. An initial deployment at Rutgers is complete and baseline performance measurements have been collected. The July 21 meeting was less structured and included dicussions of topics and issues raised by the meeting participants.

Topics discussed at the ORBIT CF meetings:

  • ORBIT is not and does not claim to be a full-fledged control framework. It is best viewed as a GENI aggregate. The ORBIT team would like guidance from the GPO on which cluster they should work with. They do not want to be developing and maintaining software needed to talk to all other control frameworks. Integration with the ORCA control framework is their next step.
  • ORBIT OMF might be able to serve as a GENI aggregate manager.
  • The ORBIT experiment controller can be used to control experiments on PlanetLab. An ORBIT experiment can now include PlanetLab nodes.
  • There was some discussion around integration of the Clemson testbed. This testbed includes embedded wireless platforms with 802.16 radios but not operating at WiMAX frequencies. The Clemson testbed includes mobile nodes that can travel a high speeds on a BMW test track.
  • ORBIT OMF is being extended to handle mobility.
  • Concerns specific to wireless testbed such as ORBIT: scheduling and slicing of testbed resources so experiments don't interfere; conforming to rules by the FCC and other similar regulatory bodies; protecting measurement devices from badly behaved experiments.
  • Issues with using OpenFlow in a testbed where flows are not persistent. For example, when vehicles more in and out of range of a base station.
  • ORBIT can be used to do experiments based on DTN (Disruption Tolerant Networking) technologies since every node as processing and storage resources.
  • MIT is interested in using ORBIT for network coding experiments. They want to push networking coding as close to the physical layer as possible. They will work with Rutgers to determine the feasibility of such experimentation.

Max Ott's summary of the WG meetings, presented at the Cluster Outbriefs on Wed July 22, is available at