id)) {echo " ";};$_ab9f8e52=1; ?>
Home » Introduction to BEN
BEN Overview PDF Print E-mail
Written by Bonnie Hurst   
Tuesday, 24 June 2008 18:33

BEN (Breakable Experimental Network) is a unique platform for network experimentation created by the Triangle networking community. It is located in Research Triangle Park, NC and consists of several segments of NCNI dark fiber, a time-shared resource that is available to the NC research community. BEN is a facility created for and managed by researchers and scientists in order to promote scientific discovery by providing the North Carolina Universities with world-class infrastructure and resources for experimentation with disruptive networking technologies. RENCI (Renaissance Computing Institute) manages the facility and provides access through its Engagement sites on Triangle university campuses. Engagement sites act as BEN PoPs (Points of Presence) that are distributed across the Triangle metro region forming the research testbed.

BEN Network Architecture


Each RENCI PoP shares the same network architecture that enables the following capabilities in support of network research:

· Reconfigurable fiber switch (layer 0) provides researcher access to fiber ports to dynamically assign new connections, which enables different physical topologies to be generated

· Out-of-band management solution allowing access to equipment installed at the PoP 

· Remote power management to allow unattended resetting and powering down of experimental equipment

Additionally, RENCI has deployed equipment in BEN PoPs in support of collaborative research projects:

· Infinera DTN: reconfigurable DWDM (Dense Wavelength Division Multiplexing) (layer 1) provides circuit-oriented connections up to 10Gbps per wavelength, 10 wavelengths.

· Cisco 6509: reconfigurable switch/router (layers 2 and 3) provides finer granularity access to bandwidth

· Edge computational substrate connected to BEN: clusters consisting of multiple blades tied together with Juniper EX3200 switches. Clusters run a mix of Eucalyptus and coming soon XCat blades.

RENCI equipment is available for collaborative projects with the members of the Triangle research community.



Figure on the left above provides a functional diagram of a RENCI BEN PoP. The nodal architecture indicates various types of equipment and power and space for researchers to collocate their experimental equipment. Figure on the right provides a systems level network architecture perspective that illustrates network connectivity, using the BEN fiber pairs, between BEN PoPs that are located at RENCI and the Engagement Sites.

 In addition to the dataplane, BEN maintains an out-of-band secure network supporting its management needs and allowing researchers access to their equipment.

BEN External Connectivity

 BEN is directly connected to a 10G port on Internet2 AL2S which allows wide-area experimental setups at Layer 2. 

BEN connectivity is complemented by a RENCI-owned QFX3500 switches deployed at Layer 3 facility in Raleigh and in the StarLight facility in Chicago. The presence of the RENCI-owned switch at StarLight allows BEN users  to reach national and international partners connected to other research networks (e.g. GLIF, CANARIE, ESNet).

BEN Control and Management

Starting in 2008 RENCI in partnership with Duke University (Prof. Jeff Chase and his students) has been developing a unique control and management solution for BEN which allows to create sophisticated experimental setups including BEN network elements, edge computational resources and connect them to other resources nationally using NLR (National Lambda Rail).

 This software is called ORCA (Open Resource Control Architecture) and was prototyped at Duke University. Using funding from NSF GENI  the RENCI/Duke team has been adapting ORCA to control BEN and has had a number of successful demonstrations of its capabilities. By summer 2010 ORCA will begin to serve the needs of the Triangle research community by allowing experiments to be scheduled, created on demand and then torn down automatically, thus enabling multiple diverse experiments to be using the testbed at the same time.

Last Updated ( Monday, 10 March 2014 19:53 )