Tuesday, January 28, 2025 - Tutorial Day
Location: SDSC Auditorium & Synthesis Center [some sessions will be recorded]

Tutorial: FPGA Applications on Nautilus

Mohammad Sada and Elham Khoda, SDSC, UC San Diego
[SDSC Synthesis Center]
This hands-on tutorial will guide participants in provisioning Xilinx Alveo U55C FPGAs on Nautilus and using Coder to manage these cards without prior Kubernetes experience. It will cover the FPGA development environment on Coder, which includes Xilinx tools like Vivado, Vitis, and Vitis AI, along with a pre-configured License Server. The tutorial will feature practical examples of FPGA applications, showcasing ML/AI workloads and scientific computing that leverage the processing capabilities of the FPGAs.

Tutorial: Title: Expanding AI/ML Coursework on Your Campus with Jupyter Notebooks Powered by NRP

Mike Farley, Kyle Krick SDSU, and Alex Hurt, U Missouri [Recorded, SDSC Auditorium]
Description: Leverage the National Research Platform (NRP) to enable and elevate interactive and hands-on AI/ML coursework at your institution with the robust capabilities of Jupyter Notebooks powered by the NRP. Discover best practices from other campuses, explore effective course development and delivery strategies, and engage with your peers in fostering collaborative support as we advance together to elevate AI/ML education.

Tutorial: Programmable Networking and Testbed Integration

Mohammad Sada, SDSC, UC San Diego, and Justas Balcas, ESnet
[SDSC Synthesis Center]
This hands-on tutorial will explore programmable networking using P4 on Nautilus, featuring P4-programmable Tofino switches and Xilinx Alveo U55C SmartNICs capable of 100 Gbps per port. Participants will learn how to run P4 workloads and provision network paths that integrate with other testbeds (e.g., FABRIC and Chameleon Cloud) using the ESnet SENSE operator. The session will also introduce Seam, a JupyterLab-based environment designed for seamless multi-testbed experimentation. JupyterLab with Selkies and Jupyter AI

Lunch

Location: SDSC Auditorium

Tutorial: Kubernetes for AI Research and Education

Mahidhar Tatineni, SDSC, UC San Diego
[Recorded, SDSC Auditorium]
In this tutorial the attendees will get an overview of the Kubernetes architecture, typical procedures for running jobs, deployment of AI services, learn how to use various storage options, interactive access including Jupyter, and run examples drawn from AI using both CPU and GPU resources. Theoretical information will be paired with hands-on sessions operating on the National Research Platform (NRP) production Kubernetes cluster which features a variety of compute (CPU,GPU) and storage resources.

Tutorial: National Data Platform Education Hub and Points of Presence

Melissa Floca and Pedro Ramonetti, UC San Diego

[Recorded, SDSC Auditorium]
This tutorial offers an introduction to using NDP to create data challenges and class projects. Existing modules for data challenges and classes will be presented and we will walk users through the experience for both educators and learners on the platform. In addition, we will provide time to show users how to set up their own point of presence on NDP. This tutorial will serve as an opportunity for researchers and educators to formally join our research team in 2025 efforts to co-develop educational materials and features and functions for NDP.

NRP Network and Server Performance Monitoring With sFlow

Neil McKee, InMon
[SDSC Synthesis Center]
Learn how open-source sFlow agents are deployed on every Nautilus node to provide detailed node and pod traffic visibility; scalable host and pod CPU/GPU performance profiling; real time analysis of dropped packets, and insight into the delay, loss, jitter, path-MTU and retransmissions seen by all TCP traffic flowing cluster-wide. Learn how to deploy similar monitoring in your own infrastructure. Will include live demos of both free and commercial tools.

Tutorial: CENIC AIR Achieved 2024 Science DMZ setups and user stats-Lessons Learned

Christopher Bruton, Tom DeFanti, UC San Diego

[Recorded, SDSC Auditorium]
The CENIC AIR FAQ, available at https://cenic.org/publications/cenic-air-faq-july-2024) will be presented, and audience questions will be encouraged. Included: a brief definition of the CENIC AI Resource (CENIC AIR), why it was created, what it consists of, and how it relates to the National Research Platform (NRP). Then CENIC AIR for research and instruction will follow, addressing questions from researchers and faculty at universities and colleges on how to access the computers and storage provided by CENIC AIR institutions. Finally, connecting your institution’s resources to CENIC AIR, addressing infrastructure-related questions about connection bandwidth, implementing a Science DMZ in proven ways, adding compute and storage resources, obtaining funding, and the roles and responsibilities of your institution, CENIC, and the National Research Platform regarding design, implementation, maintenance, and support.

Wednesday, January 29, 2025
Location: Qualcomm Institute (QI) Auditorium [all sessions recorded]

The State of NRP

Frank Wuerthwein, Director of San Diego Supercomputer Center, UC San Diego
After a brief description of the vision behind NRP, we will discuss progress within the last year towards that vision, and put the program of this workshop in context.

Data on NRP, Challenges and Opportunities

Moderator: Frank Wuerthwein, SDSC, UC San Diego

- Derek Weitzel, University of Nebraska-Lincoln: Storage options supported on NRP today (20min)
- Panel on how researchers use data on NRP today (20min)
- Panel discussion and Q&A with audience on Data on NRP (20min)

Lunch
Location: QI Courtyard

Optional Session: Bring your lunch into VROOM for presentations starting at 12:10 PM
Harnessing the Power of JupyterLite to Bridge the Digital Divide
Moderator: Rick Wagner

- Dhusdee Chandswang, CSU Fullerton

- Kyle Krick, SDSU

- Willie Peng, CSU Fullerton

- Ben Smarr, UC San Diego

- Robert Twomey, UC San Diego

- Dung Vu, CSU San Bernardino

Internet2 and QUILT initiatives with Regional Networks

Moderator: Larry Smarr, UC San Diego

James Deaton, Internet2 and Jen Leasure, The Quilt
Two additional Regional Networks are developing in the CENIC AIR model–GPN AIR and SoX AIR. The Quilt and Internet2 will use their meetings of member organizations to further propagate the AIR model for Regional Networks in 2025.

Building User Bases: Examples from CC* Awardees;

Alex Hurt, University of Missouri and Derek Weitzel, University of Nebraska-Lincoln
The National Research Platform (NRP) offers a wealth of resources and expertise to researchers and educators alike. However, expanding its user base within academic institutions remains a significant challenge due to both technical and logistical hurdles. Alex Hurt and Derek Weitzel will discuss these challenges in detail. They will explore the specific technical barriers and logistical difficulties that hinder NRP adoption. Additionally, they will share their approaches to overcome these obstacles and foster a growing NRP user community on their campuses and in their respective regions. Examples from exemplar CC* awardees like SDSU, UCR, MU, UN-L, ASU, others.

Addressing Classroom Use Cases for Broader Impact (Panel)
Moderator: Lauren Michael, Consultant, Internet2 Research Engagement

Rosanna Overholser, CSU Poly Tech University Humbolt, and  Jeho Park, Claremont McKenna College
While national-scale cyberinfrastructure providers tend to operate out of R1 or R2 institutions, less than 10% of the institutions of higher education and just 6 of the 20 million undergraduate students across the country are represented by these campuses. To meet national demand for educating future generations in data science and computational skills, cyberinfrastructure providers must prepare to engage more broadly than ever before, with a need to understand education use cases outside the walls and resources of campuses similar to their own. Beyond their roles in educating a much larger portion of the nation’s future workforce, smaller and lesser-resourced institutions bring an agile, efficient, and necessity-driven approach to integrating computing and data science in the classroom. With institutional missions centered significantly around undergraduate retention, recruitment, and job placement, many of these institutions also specialize in student experiences that empower underserved minorities and nontraditional students. This panel will further explore the educational strengths of these institutions, their current needs and strategies for computational education, and how cyberinfrastructure providers might support their technology needs in the numbers necessary to achieve a truly national impact.

Advancing AI and Technical Integration for California Community Colleges: A Collaborative Approach

Peter Maharaj, Associate Vice Chancellor, Information Technology Services.
This session highlights the innovative partnership between the San Diego Community College District (SDCCD), Riverside Community College District (RCCD), Los Angeles Community College District (LACCD), UCSD, SDSU, and CENIC to enhance curriculum development, improve access, and implement cutting-edge technical advancements. The collaboration focuses on leveraging the National Research Platform (NRP) to democratize artificial intelligence tools and foster equitable educational opportunities across California’s 73 community college districts.

Participants will gain insights into

- Curriculum Development. Strategies for integrating AI into existing and new coursework to prepare students for emerging career opportunities.
- Access and Equity. Initiatives aimed at bridging the digital divide and enabling students and faculty to engage with advanced technologies fully.-
- Technical Enhancements. Implementation and expansion of a robust infrastructure, including cloud environments and AI-driven solutions, to support scalable and sustainable innovation.
 

This session will also share the vision of aligning community colleges with national initiatives, such as the National AI Research Resource (NAIRR) and National Science Foundation (NSF) CC*, to position California as a leader in technology-enabled education. Attendees will leave with knowledge and actionable takeaways for fostering collaboration and driving meaningful change within their institutions.

AI Education: Expanding the AI Curricula

Moderator: Mahidhar Tatineni, SDSC, UC San Diego (Panel)
Carl Boettiger, and Eric Van Dusen, UC Berkeley; Ashley Atkins and Mary Thomas, SDSC, UC San Diego
This panel showcases perspectives on integrating AI into education, with an emphasis on how existing resources, such as the National Research Platform (NRP), are being innovatively leveraged for use in classes. Discussions will be focused on identifying key needs for future advancement.

Navigating the AI Disruption at Scale: Perspectives and Strategy from the CSU Chancellor's Office
Kendra Ard, Chief Infrastructure Officer for Infrastructure, Support & Communications Services, CSU Chancellor’s Office
Matt Hughes, Systemwide Infrastructure Services Network Architect, CSU Chancellor’s Office
Generative AI has emerged as a powerful yet disruptive force in higher education, reshaping how we teach, learn, and work. As AI capabilities rapidly evolve, educators, administrators, and students are all in a state of discovery, collectively navigating the new opportunities and challenges this technology presents. In this session, we will explore the collaborative journey we are on as we develop responsive strategies at the enormous scale of the nation’s largest four-year higher education institution.

The California State University Chancellor's Office is in the early stages of designing an AI strategy that aims to guide the system through this transformation. This strategy is a work in progress, informed by the CSU Generative AI Committee recommendations, the Chancellor's evolving priorities for the institution, and partnerships with the State. It is intended to provide a roadmap toward equitable access across campuses, continuous training and support, and enabling governance structures, while also leaving room for adaptation, experimentation, and localized innovation. This presentation will share preliminary insights from ongoing efforts and discuss both the challenges and successes we’ve encountered so far.

Social Event sponsorsed by Ciena

Akbar Kara, Ciena, Host
Location: QI Atkinson Lobby and Courtyard

Thursday, January 30, 2025
Location: Qualcomm Institute (QI) Auditorium [all sessions recorded]

Keynote: Promises and Challenges toward Clean Energy From Fusion: AI and Data Analytics Approaches

Raffi Nazikian, General Atomics
The quest for clean energy from nuclear fusion has received a big boost within the last year by the first ever fusion experiment that produced more energy from the plasma than it required to produce the plasma. This leads to global excitement over fusion energy as a viable long term candidate for clean, carbon-free, ubiquitous, and cheap energy. We will review some of the challenges that remain, and how AI is expected to become an accelerant to progress towards a commercially viable fusion reactor within the next decade or two.

AI/ML Computing Across Scientific Disciplines

Larry Smarr, UC San Diego
NRP storage and computing resources are being used to investigate how AI/ML algorithms can be applied to a wide range of disciplines. Examples of such Nautilus research namespace projects are described from astronomy, biology, chemistry, Earth & environmental sciences, genomics, healthcare, and physics.

Equitable AI Alliance

David Goldberg, SDSU
As artificial intelligence (AI) becomes integral to the workforce, unequal access to AI resources threatens to widen existing gaps, particularly for students from historically underrepresented communities. Recognizing the need for equitable AI education, San Diego State University (SDSU), UC San Diego (UCSD), and the San Diego Community College District (SDCCD) have established the “Equitable AI Alliance”. The Alliance is built on three key activities: (1) AI needs assessment, (2) AI literacy education, and (3) AI access. It begins with a comprehensive survey to understand AI-related perspectives and needs across campuses (AI needs assessment), which will inform the development of AI literacy programs, including micro-credential courses for students, faculty, and staff (AI literacy education). Additionally, the initiative will ensure equitable access to AI tools through the construction and expansion of high-quality internal AI tools that all three institutions will leverage to bridge the digital divide (AI access). The Equitable AI Alliance aims to provide every student, regardless of their point of entry into California’s higher education system, an accessible onramp to AI skills-building. This scalable model for AI education promotes equity and career readiness across California’s higher education landscape and sets a precedent for national and international efforts.

Lunch

Location: QI Courtyard

Optional Session: Bring your lunch into VROOM for presentations starting at 12:00 PM
International Extensions to NRP

Maxine Brown, UIC, Panel Chair
Panel speakers present lightning talks from several international Research Platforms that are extending NRP capabilities across the globe. In this era of Big Data and Artificial Intelligence, these Platforms reduce technological divisions and geographical boundaries among services and resources to enable seamless use by data-intensive applications, and provide scientific communities with efficient access to the specific resources they require. Throughout the year, the people developing these Platforms work together to design, implement, and test strategies for next-generation, distributed, network services and infrastructure, and to share lessons learned with others through demonstrations at major conferences and workshops, publications, and presentations, including the annual Global Research Platform (GRP) Workshop. Ultimately, Research Platform advancements in architectures, services, and technologies are disseminated and put into operational practice within National Research & Education Networks, Open Exchanges and network testbeds around the world for the benefit of the broader Research & Education communities.
Panelists:

- Joe Mambretti, StarLight/iCAIR/NU — The Global Research Platform

-  Julio Ibarra, FIU — AmLight and AMPATH: Supporting International Research Platforms for Latin America, the Caribbean, and Africa

- Jeonghoon Moon, KREONET Center, KISTI — Asia Pacific Research Platform and Related Work

- Brenna Meade, Indiana University — TransPAC

- Harvey Newman, Caltech — GNA-DIS: A Next-Generation System for Data-Intensive Sciences: Progress, Outlook and Results from SC24

- Cees de Laat, University of Amsterdam (UvA)— Cyber Infrastructure Challenges in the Roaring Twenties

Advanced Technology session: New Compute Architectures and FPGA software environments on NRP

Mohammad Sada, SDSC, UC San Diego
We are supporting multiple different programming environments for the FPGAs on NRP, as well as other unusual compute architectures. In this talk we discuss what’s available, with a focus on what’s new since last year.

SmartNICs--What’s Working and What’s Next? ESnet, CERN Successes; AutoGOLE/SENSE Orchestration
Moderator: Tom Lehman, ESnet
This panel will explore the current landscape of SmartNICs and FPGAs on Nautilus, highlighting successes from key projects involving SC24, ESnet, and experiments with CERN. The discussion will delve into interoperability with federated testbeds such as FABRIC and Chameleon Cloud, focusing on dynamic Layer 2 and Layer 3 network multi-domain service provisioning with AutoGOLE/SENSE orchestration and end-to-end network services. Panelists will discuss both the technical achievements and challenges faced in integrating SmartNICs into high-performance network environments, and share insights into what the future holds for SmartNICs.

Panelists:

- Justas Balcas, ESnet
- Joe Mambretti, Northwestern University
- Harvey Newman, Caltech
- Mohammad Sada, SDSC, UC San Diego

The National Data Platform (NDP) as a Research and Education Resource for AI

Manish Parashar, University of Utah
Open and equitable access to scientific data is essential to addressing important scientific and societal grand challenges, and to research enterprise more broadly. However, achieving scientific and societal scale goals requires educating and enabling a new generation to participate in AI. This session discusses the importance and urgency of open and equitable data access and introduces the vision and architecture of the National Data Platform, a project aimed at catalyzing an open, equitable and extensible data and service ecosystem for AI workflows in research and education.

Arts, Computational Media & AI

Robert Twomey, UC San Diego
Access to affordable GPU compute resources has become a necessity for contemporary work in arts and computational media. This talk presents case studies where NRP resources have enabled generative media research and education: supporting an AI film festival, experimental human-robot interaction, and AI theater performance.

Final session: Open Mic

Moderator: Frank Wuerthwein, SDSC, UC San Diego