= Control Framework Working Group = Chairs: Jeff Chase, Duke University; Rob Ricci, University of Utah[[br]] System Engineer: Aaron Falk, GENI Project Office '''Agenda''' TOPIC 1: GENI APIs (60 minutes) The first part of the Control Framework Working Group meeting at GEC9 will consist of a discussion of the GENI APIs that are under development: http://groups.geni.net/geni/wiki/GeniApi This discussion will kick off with a talk ([attachment:TMitchell-Pragmatic.pdf slides]) by Tom Mitchell of BBN, who has been leading the design and a few different implementations of these APIs. We would like to follow this up with short (5 minute max) talks by members of the working group, recommending changes (small and large) to the API, or discussing the need (or lack thereof) for a common API. Talks proposing changes should include a) the deficiency or opportunity motivating the change, b) the specifics of the change relative to the current API, and c) the benefits of the proposal. If you're interested in giving one of these talks, please email the WG chairs (myself and Jeff Chase) and our SE, Aaron Falk. We want to leave time for discussion of these ideas, and time is limited, so we may need to limit the number of talks that we accept. Confirmed short talks: * Ted Faber, ISI ([attachment:cf_slides.pdf slides]) * Thierry Parmentelat, INRIA * Anirban Mandal, RENCI ([attachment:GEC9-CFWG-Anirban.pptx slides]) * Tom Mitchell, BBN ([attachment:TMitchell-Changes.pdf slides]) * Justin Cappos, Washington ([attachment:Cappos_CWFG_GEC9.ppt slides]) TOPIC 2: Talks/discussion about next directions for GENI API or architecture in general. (45 minutes) * Neuca, Ilia Baldine, ''RENCI'' ([attachment:GEC9-NEuca.pptx slides]) * Stitching, Tom Lehman, ''USC ISI'' ([attachment:2010-11-03-network-stitching-lehman-control-wg-gec9.ppt slides]) * Expedient: OpenFlow Client-GUI/Clearinghouse, Jad Naous, ''Stanford'' ([attachment:'Expedient and Unified APIs.pdf' slides)] TOPIC 3: Invited talk (15 min) Speaker: Mathieu Lacage of INRIA's Planete team ''In this talk, we will present NEPI, an experiment-plane tool which attempts to provide a unified abstraction on top of existing testbeds to describe the network-level and application-level aspects of a network experiment and automate resource reservation, application deployment, experiment monitoring, and trace collection. [http://ralyx.inrialpes.fr/2009/Raweb/planete/uid22.html NEPI project page]'' ([attachment:geni-nepi.pdf slides])