PEARC 2026
- Minneapolis, MN
Stop by the Globus booth to learn what’s new. Join one of the following sessions:
Globus Compute for Administrators: Enabling Streamlined Multi-user Function Serving on Federated CI
Increasing data volumes and hardware heterogeneity are driving scientific workloads toward distributed execution across diverse resources. Globus Compute is a hybrid-cloud Function-as-a-Service (FaaS) platform that enables researchers to execute functions close to data, exploit specialized hardware, and scale beyond individual systems. Users provide only a function and its inputs, while the platform manages resource provisioning, remote execution, monitoring, and asynchronous result delivery. This tutorial focuses on deploying and administering Globus Compute end points, which turn existing resources—ranging from laptops to supercomputers—into secure FaaS endpoints. Topics include mutli-user deployments, administrator controls, policy enforcement, and the role of Globus Compute in modern cyberinfrastructure, such as NSF ACCESS.
Session Prerequisites: Intermediate
Target Audience: Administrators of computing resources (HPC systems, cloud platforms)
Required Participant Resources: Participants will need to bring a laptop running any OS with the ability to either run Docker containers locally or connect via SSH to a hosted cloud VM. Basic programming experience (ideally with Python) and basic Linux experience are preferable.
Building Scalable Agentic Systems for Science: Concepts, Architectures, and Hands-On with Academy
Agentic systems, in which autonomous agents collaborate tos olve complex problems, are emerging as a transformative methodology in AI. However, adapting agentic architectures to scientific cyberinfrastructure—spanning HPC systems, experimental facilities, and federated datarepositories—introduces new technical challenges. In this half-day tutorial, we introduce participants to the design, deployment, and management of scalable agentic systems for scientific discovery. We will present Academy, aPython-based middleware platform built to support agentic workflows across heterogeneous research environments.Participants will learn core agentic system concepts, including asynchronous execution models, stateful agent orchestration, and dynamic resource management. A guided hands-on session will help attendees build and launch their own agentic workflows. We will present case studies in materials discovery, biology, and chemistry. This tutorial is designed for researchers, developers, and cyberinfrastructure professionals interested in advancingAI-driven science with next-generation autonomous systems.
Session Prerequisites: Novice
Target Audience: Researchers, students, developers, and practitioners
Required Participant Resources: Attendees will need to bring a laptop with a working Python environment. Mac, Linux, and Windows Linux Subsystem environments will be supported and tested. We will provide a Docker container image from which attendees can conduct the hands-on exercises if they do not have a suitable local environment. Participants will not need access to any external computing resources.
Streamlining Collaboration in Protected Data Environments Using Globus
Data-intensive research increasingly depends on sharing large and sensitive datasets across institutional boundaries, including movement from institutional storage systems and shared facilities (campus clusters, core facilities, national labs, cloud platforms) to external collaborators and analysis environments. Ensuring that these data egress workflows are secure, compliant with institutional and regulatory policies, and repeatable is a major challenge for research computing, campus IT, and data stewards.This tutorial introduces attendees to Globus capabilities for secure, policy-aware data sharing and controlled data egress, with a particular focus on automation using Globus Flows. We will demonstrate how to use Globus to:- Enable institutional storage for secure, compliant access via Globus.- Implement access control and sharing policies via Globus mapped and guest collections to enable secure data egress to external collaborators and environments (including cloud). Automate complex, multi-step data egress workflows using Globus Flows. Log, monitor, and audit data movement for compliance and reporting. The tutorial combines short lectures with hands-on exercises. Participants will create collections, configure sharing and access controls, conduct secure transfers, and automate common egress workflows (e.g., staging, validation, and delivery) using Globus Flows. We will also discuss institutional patterns and reference architectures for data egress, including landing zones, controlled sharing enclaves, and automated “pipeline” approaches that embed policy into research workflows. The tutorial is intended for research computing professionals, system administrators, security and compliance staff, and research data managers responsible for enabling and governing secure, automated data sharing across institutional boundaries.
Session Prerequisites: General
Target Audience: Research computing and data (RCD) professionals, System administrators for institutional storage, clusters, and shared facilities, Campus IT security and compliance staff, Research data managers and data stewards, PIs and lab managers overseeing data-intensive projects with sharing/compliance requirements.
Required Participant Resources: Laptop with modern web browser (Chrome, Firefox, Edge, or similar). Optional: SSH terminal software (bash, iTerm, PuTTY, or similar).