Project 1766 - PrimoGENI Spiral 2 Year-end Project Review

1. Project 1766 - PrimoGENISpiral 2 Year-end Project Review Florida International University PI: Jason Liu; Co-PIs: Julio Ibarra, Heidi Alvarez Technical Staff:Ernesto RubiGraduate Students: Miguel Erazo, Nathanael Van Vorst, Undergraduate Students: Eduardo Tibau, Eduardo Pena 8-30-2010

2. Project Summary • The goal of PrimoGENI is to incorporate real-time network simulation capabilities into the GENI "ecosystem". We are extending our existing real-time network simulator PRIMEX to become part of the GENI federation. • PrimoGENI will support large-scalehybrid GENI experiments involving physical, simulated and emulated network entities. • A PrimoGENI experiment will contain millions of simulated network entities (hosts, routers and links) and thousands of emulated elements running unmodified network protocols and applications. • Core activities include: • Development and prototyping of GENI aggregate: augment PRIMEX with GENI interface, run under ProtoGENI control framework, exploit full experiment control • Infrastructure development: connect geographically distributed resources nationally and internationally • Early federation experiments and opt-in support capabilities: enable experiments of real applications across various GENI resources • Education and outreach: organize tutorials, lectures, and demonstrations 8-30-2010

3. Milestone & QSR Status 8-30-2010

4. Accomplishments 1: Advancing GENI Spiral 2 Goals • Design and Develop PrimoGENI Aggregate Manager: • Set up PrimoGENI environment on top of ProtoGENI cluster • Fully integrated with ProtoGENI control framework: augment ProtoGENI control flow with customized scripts for PrimoGENI experiments • PrimoGENI 2.0 design: adopt a decentralized database-focused approach with client IDE for designing models, launching experiments, and monitoring results • Enhanced ProtoGENI nodes with customized OpenVZ image supporting PrimoGENI experiments within slices • Re-engineer PRIMEX: • Streamlined design of PRIMEX based on “network scripting”: to manage the entire ``life cycle'' of a large-scale network experiment, including model construction, configuration, resource specification, simulation deployment, execution, online monitoring and control, experiment data collection, inspection, visualization and analysis • Develop large scale network experiment support (e.g., spherical routing), for performance and memory efficiency • Developing high-speed conduit between virtual machines for hybrid experiments • Client API: Slingshot • Design IDE for PrimoGENI experiments, including model development, visualization, deployment, monitoring and control 8-30-2010

5. Accomplishments 2:Other Project Accomplishments • Two successful demos at GEC 7 and GEC 8 • Julio Ibarra delivered plenary talk at GEC 8 on extending GENI to Brazil and internationally • Collaborating with the UML team to develop a high performance conduit for simulation and emulation on OpenVZ • Planning with UML team for an end-to-end inter-campus VLAN test and large-scale network experiments • Planning with Brazil partners (Kyateraand GIGA) for international connectivity running PrimoGENI and OpenFlow 8-30-2010

6. Issues • PrimoGENI development turns out to be a substantial undertaking. We have gone nearly 9 months into the project and have so far been short staffed to deliver a fully functional prototype in time. We were less than two months late for delivering PrimoGENI 1.0. But we finally managed. • For every GEC so far, we sent at least three people to take care of the demonstrations. Major funding for travel has been from other sources. As PrimoGENI becomes more mature, we plan to carry out several outreach activities and certainly more demonstrations, in spite of the shortage in such funding. 8-30-2010

7. Plans • What are you plans for the remainder of Spiral 2? • We will roll out PrimoGENI 2.0 design in a week (as a matter of fact, we are well on our way to implement PrimoGENI 2.0) • Design and implement a network experiment using PrimoGENI showcasing its capabilities (hopefully at the next GEC venue) • Establish inter-campus VLAN connectivity with UML and design realistic experiments using both UMLPEN and PrimoGENI aggregates • The GPO is starting to formulate goals for Spiral 3. What are your thoughts regarding potential Spiral 3 work? • Establish PrimoGENI aggregates through international connections (in particular, between US and Brazil) • PrimoGENI will run on multiple GENI aggregates that are geographically distributed and possibly across different control frameworks (for end-to-end slicing, e.g., across GENI racks and OpenFlow switches) for multiple simultaneous large-scale high-fidelity network experiments • We will expand outreach and education activities. We will collaborate with large education and training projects, offering education training materials, preparing lectures, demonstrations and tutorials, in return for venues to disseminate our work 8-30-2010