Serverless Research Data Portals Using the Globus Platform - Tutorial

Serverless Research Data Portals Using the Globus Platform R. Lee Liming Globus University of Chicago Chicago, U.S.A. [email protected]g Joe Bottigliero Globus University of Chicago Chicago, U.S.A. [email protected]g Andrew Boughton Globus University of Chicago Chicago, U.S.A. [email protected]g Abstract—This 180-minute, interactive tutorial at the Gateways 2025 conference introduces a way to create research data portals without requiring new infrastructure. Participants will practice deploying serverless data portals with a variety of features. Participants will build data collections, construct searchable metadata, and construct data ingestion pipelines using the Globus platform. Index Terms—Big Data, Data dissemination, Indexes, Metadata, Platform as a service, Portals, Serverless computing I. INTRODUCTION Making research data FAIR (Findable, Accessible, Interoperable, and Reproducible) is an increasingly important requirement for research teams. [1] Research data portals are web applications that enable researchers to make their data accessible to other researchers and the public. They focus on data discovery and accessibility and assist with interoperability and reproducibility. [2] But data portals are not widely available to teams in all fields and institutions, particularly for teams that produce large scales of data. Designing, operating, and maintaining these portals may require significant effort, even after the project has ended. In a recent paper [3], we introduced ZRDP: Zero Code and Infrastructure Research Data Portals, a way to build and deploy a research data portal without writing code and without operating servers. Globus has created templates and best practices for applying the concepts introduced in ZRDP. The resulting serverless data portals use institutional data storage and identity services and leverage cloud-based Globus platform services for interactive features such as data transfer and search. This 180-minute tutorial, taught by experienced data portal practitioners, walks participants through the process of creating serverless data portals, beginning with a simple 10-minute bare-bones portal and progressively adding features. The final example is a serverless data portal that enables access to datasets housed in institutional data storage, including a search index, highspeed Globus data transfer, and a web interface for adding new datasets. II. TARGET AUDIENCE AND REQUIRED EQUIPMENT There are two target audiences for this tutorial: •Researchers who need to share their data products (physical observations, instrument or simulation outputs, aggregated datasets, model training datasets, etc.) with research colleagues or with the public •Research IT and library personnel who set up data portals for the researchers they support The tutorial is of particular relevance to researchers, librarians, and research IT personnel who do not (or prefer not to) operate their own servers. Skill levels: Any level (Beginner to Advanced) To follow this tutorial, we recommend two screens: one to watch the presentation, and the other for the hands-on examples. Participants should have a web browser, internet connection, keyboard, pointing device, and a large enough display to read the example text. A GitHub account is recommended. III. TUTORIAL FORMAT AND AGENDA This 180-minute, interactive tutorial will consist of brief presentations followed by longer interactive, hands-on examples. 1) Introduction: The Modern Research Data Portal (MRDP) design pattern (30 min) 2) Serverless data portal in 10 minutes (10 min) 3) Make your data accessible - data storage, downloads, and transfers (20 min) 4) Make your data findable - metadata and indexing (30 min) 5) BREAK (30 min) 6) Serverless search portal (30 min) 7) Add data to your portal - uploads and ingests (60 min) IV. TUTORIAL CONTENT A. The Modern Research Data Portal Design Pattern We begin by introducing the Modern Research Data Portal design pattern and the benefits it offers to research teams and research institutions. This includes a brief overview of the Globus platform, including: software for connecting data storage and compute systems, federated accounts using InCommon and eduGAIN, secure data transfer, secure remote compute, indexed search, and reliable workflow execution. B. Serverless Data Portal in 10 Minutes For the first interactive section of the tutorial, we’ll show how to quickly set up a serverless data portal on GitHub Pages that enables others to transfer, download, and link to your data. We’ll provide a sample data collection, but participants can substitute their own if they have one. C. Make your data accessible We’ll show how to access institutional data storage via Globus; configure a guest collection that makes research data in shared storage accessible via a data portal; add and organize research data; and transfer, download, and link to data with a serverless data portal. This section begins with a brief presentation and then is largely interactive. D. Make your data findable For research teams and research computing centers with many datasets, we’ll show: how to create and extract metadata to make data findable; how to create a search index and add entries for datasets; and how Globus Search’s search interfaces work. This section is a balance of presentation and interactive examples. E. Serverless search portal After setting up a search index, we’ll configure and deploy a second serverless data portal that includes a search interface. This portal, hosted on GitHub Pages, enables data discovery, download, and transfer. F. Add data to your portal The final section of the tutorial is for teams who need to enable team members (or others) to add datasets to their data portal. With a balance of presentation and interactive examples, we’ll show how to create an input form for manually entering a dataset description; set up a Globus flow that transfers data to institutional storage and make the dataset searchable; configure permissions on the flow to prevent unauthorized use; and configure the data storage to run the flow when someone attempts to transfer data to it. V. PRESENTERS Lee Liming is the Director of Professional Services (Globus) at the University of Chicago. In 20+ years with the Globus team, Lee has contributed to national research initiatives including computing (NSF’s TeraGrid, XSEDE, ACCESS), bioinformatics (NIH’s BIRN, CFDE), and climate (DOE’s ESGF2-US). He has presented tutorials at SCxy, PEARC, Gateways, US-RSE, and GlobusWorld conferences, and has worked at the University of Michigan, ProQuest LLC, Argonne National Laboratory, and the University of Chicago. Andy Boughton is a Professional Services Engineer (Globus) at the University of Chicago. He has worked on a number of projects enabling large scale data collection and analysis across the research lifecycle. He has helped build solutions in diverse fields including chemistry, vehicle engineering, social psychology, bioinformatics, and human genetics. Example projects include bioinformatics tools such as Genes for Good and the TOPMed imputation server, as well as data portals such as the Open Science Framework / OSF Preprints service and my.locuszoom.org. REFERENCES [1] C. L. Borgman, “The conundrum of sharing research data,” Journal of the American Society for Information Science and Technology, vol. 63, no. 6, pp. 1059–1078, 2012. [Online]. Available: https://onlinelibrary.wiley.com/doi/abs/10.1002/asi.22634 [2] K. Chard, E. Dart, I. Foster, D. Shifflett, S. Tuecke, and J. Williams, “The modern research data portal: A design pattern for networked, dataintensive science,” PeerJ Computer Science, vol. 4, p. e144, 01 2018. [3] J. Bottigliero, R. Anathakrishnan, K. Chard, R. Chard, and I. Foster, “Zero code and infrastructure research data portals,” in Practice and Experience in Advanced Research Computing 2024: Human Powered Computing, ser. PEARC ’24. New York, NY, USA: Association for Computing Machinery, 2024. [Online]. Available: https://doi.org/10.1145/3626203.3670595