In late May, the National Science Foundation, through its directorate on Computer and Information Science and Engineering (CISE), and BBN Technologies announced a cooperative agreement for BBN to operate the GENI Project Office. This announcement came a few months after the announcement of the GENI Science Council (operating under the auspices of the CRA’s Computing Community Consortium). Together the announcements represent a significant step forward in the GENI’s evolution from an idea to reality. In this article we briefly summarize the current state of GENI activities and their expected evolution over the coming year.
What is GENI?
The Global Environment for Network Innovation (GENI) is an experimental facility intended to support research in a wide variety of areas including communications, networking, distributed systems, cyber-security, networked services and networked applications. The goal of GENI is to enable researchers to experiment with radical network and system designs in a way that is far more realistic than any alternative available today.
GENI seeks to serve researchers in as wide a range of disciplines as possible. People with research needs that GENI could address should feel free to contact either of the authors (ewz[at]cc.gatech.edu or celliott[at]bbn.com).
What is the GENI Project Office (GPO)?
The NSF regularly provides funding not simply for research but also for scientific infrastructure. Infrastructure requiring large investments from NSF falls under the auspices of the Major Research and Equipment Facilities Construction (MREFC) account.
Under the current plans, GENI will be built as an MREFC project. MREFC projects all go through a number of steps to show that the concept for the facility is sound, mature, properly engineered and likely to produce innovative science results. One of the early steps in the process is creating a Project Office, which works closely with NSF to shepherd the facility’s design and engineering plans through MREFC funding qualification steps. Thus, creating the GPO means that GENI is starting to move from a detailed concept to a nuts and bolts (or in this case, bits and photons) piece of research infrastructure.
The main job of the GPO over the next few years is engineering and risk reduction. The GPO is responsible for assembling a complete engineering plan for how to build GENI. Among other things, the plan must include detailed costing, a discussion of which technologies will be used, and a detailed demonstration that the infrastructure built will support the intended research. Because many parts of GENI will be novel or one-of-a-kind technologies, creating a credible engineering plan means driving down construction risks. To that end, the GPO will be awarding approximately $7.5M per year in risk-reduction projects to academic and industrial teams whose job it is to propose and test design alternatives, to show that key parts of GENI can be built, and to show how different versions of GENI will meet researchers’ needs.
Specification and design of the GENI facility will take place collaboratively within several working groups. This intense activity is open to anyone interested in developing or using the technologies or services provided by the GENI experimental platform. The GPO is committed to an open design and evaluation process, where working group recommendations come about through the rough consensus developed over the course of on-line and in-person discussions. Working group participants will be collaborating in areas such as: network substrate; resource control framework; experiment workflow and services; end-user opt-in; and operations, management, security and integration. New working groups may be created or old ones dissolved as the GENI design advances. The GPO is actively seeking working group participants and chairs. Some financial support may be available for these positions. (For information on participating in a working group, contact Aaron Falk of the GPO at falk[at]geni.net).
The GPO will be hosting engineering conferences three times a year where progress on GENI’s design will be reported, the working groups will meet, and progress and results from the GPO-funded risk reduction efforts will be reported. Participation at the conferences is open to anyone who wishes to attend.
Another important GPO role is outreach. GENI should enable groundbreaking research, and it is important to ensure that the chance to do that research, and to learn from and build upon that research, is shared throughout the research and education community and society. To that end, the GPO has a vigorous outreach program that seeks to bring various communities in contact with GENI, and enables people at all levels of the education system (from elementary school to graduate school) and from all backgrounds (geographic and ethnic) to participate. Much of this effort involves personal visits to conferences, campuses, and industry. There is also an active effort to encourage participation via internships at the GPO and with GPO contractors, through travel grants to the GENI Engineering Conference, and by using tools such as electronic newsletters.
Chip Elliott leads the GPO. Chip is Chief Engineer for Network Systems at BBN Technologies and Project Director for the GPO. He has nearly 30 years of experience leading large and technically challenging projects. His forte is taking exciting, often nascent, concepts and building them up to working practice. Most recently, Chip led the effort to build the first operational quantum key distribution network.
Assisting Chip is a team of roughly half a dozen senior people, including Heidi Picher Dempsey (Operations and Integration Manager), Aaron Falk (Interim Architect), Roscoe Giles of Boston University (Outreach Consultant), Craig Partridge (Outreach Director), Kristin Rauschenbach (Optics and Substrate Manager) and Henry Yeh (Project Manager). The team has over a hundred years of experience in research in communications systems and operations of experimental networks.
The GENI Science Council (GSC)
The GSC’s primary mission is to articulate a compelling rationale for GENI in the form of a comprehensive Research and Education Plan. This will describe the scientific and engineering research questions that GENI will make it possible to address and the educational opportunities that GENI will afford. The GSC represents the interests of the national GENI-relevant research and education community in the GENI facility, serves as the community’s interface to the GPO and, together with the GPO, ensures that all stakeholders have clear and accurate information regarding all aspects of the GENI project. The work of the GSC is carried out in a set of working committees that include Research and Education Plan, Outreach, Industrial Interaction, and Facility Architecture. GSC committee members come from both inside and outside the GSC.
Scott Shenker and Ellen Zegura co-chair the GSC. Scott is a Professor of Computer Science at UC Berkeley. He is also head of the Networking Group at the International Computer Science Institute (ICSI). Prior to joining ICSI, Scott spent 14 years at Xerox PARC. He is a past recipient of the ACM SIGCOMM Award, which recognizes lifetime achievement in the area of communication networks. Ellen is a Professor and Chair of Computer Science at Georgia Tech, where she has been on the faculty for 14 years. She is a past Editor-in-Chief of IEEE/ACM Transactions on Networking.
The current GSC membership (http://www.cra.org/ccc/participants.html#gsc) includes fifteen people from industry, research labs and academia, spanning fields ranging from technology policy to optical networking to distributed systems.
The Coming Year
So what’s up for GENI in the coming year?
The first GENI Engineering Conference will be held October 9-11 at the University of Minnesota in Minneapolis.
The first GPO solicitation is expected to come out in late 2007, with contracts awarded in early 2008. The expectation is that by the GENI Engineering Conference in the summer of 2008, we’ll begin to see the early fruits of those efforts.
The GPO expects to begin recruiting interns for the summer of 2008 shortly.
More information about the GENI Science Council and the GENI Project Office is available on the GENI web site (www.geni.net). This website is also where details about the conference, GPO solicitations and internships will appear.
Chip Elliott, Director of the GENI Project Office, is Chief Engineer for Network Systems at BBN Technologies.
Ellen Zegura, who co-chairs the GENI Science Council, is a Professor and Chair of Computer Science at Georgia Tech.
This work was supported in part by NSF Cooperative Support Agreement CNS-0714770. The views expressed in this article do not necessarily represent the views of the National Science Foundation.