Early distributed computing demonstration • 1997 : Globus Project established • 1998 : First Globus software released • 2000 : Globus Toolkit v1.0 • 2003 : Grid computing era – global adoption • 2005 : Innovations: GridFTP, data management • 2010 : Transition to SaaS architecture begins • 2011 : Globus Online (SaaS vs. developer tools) • 2014 : Rebranded as "Globus", service expansion • 2016 : Major new features beyond data transfer • 2021 : Widespread adoption • 2024 : Leading SaaS solution
to computing 8 “if mechanisms are in place to allow reliable, transparent, and instantaneous access to high-end resources, then it is as if those resources are devoted to them” (The Grid, Chapter 2) Ian Foster Carl Kesselman Steve Tuecke
Peace, 2007 Earth System Grid enables sharing of simulation outputs Discovery of Higgs Boson: Physics, 2013 “only possible because of the extraordinary achievements of … grid computing”—Rolf Heuer, CERN DG Detection of gravitational waves: Physics, 2017 LIGO scientific collaboration uses grid technologies to pool data and computing
1 Instrument, Lab server Compute Facility Set up secure tunnel 2 Globally accessible multi-tenant service • Secure tunnel across wide area networks • Leverages institutional security deployment • No changes required on the application Monitoring and notification 3 19 Stream data 3
easy-to-use platform for handling our large data uploads” - Tyson Foster, Data and Technology National Leader, National Transport Research (NTRO) Seamlessly Transfer Survey Data from Road to Cloud www.aarnet.edu.au/national-transport-research-organisation-transforms-road-surveys-with-globus
Federated access across distinct security models • Management of limits and other storage system constraints • Open ecosystem via Community Connector Program Unified data access Extensible ecosystem
molecular & cellular structure Proteins & protein families Molecular systems Data Resources Distribute PBs monthly using Globus Large scale secure data distribution
and accesses shared files; no local account required; download via Globus 2 On-prem or public cloud storage Select files to share, select user or group, and set access permissions 1 Globally accessible multi-tenant service Globus controls access to shared files on existing storage • Fine-grained access control “overlay” on storage system • Share with any identity, email, group • No need to stage data just for sharing • Time restricted sharing v 24 Compute/Storage Facility/Laptop
be run on compute endpoints 1 Globally accessible multi-tenant service Laptop, server, compute facility Compute Facility on prem or Cloud • Fire and forget function execution • Federated authentication, and local access control • Uniform interface to various compute resources • Support use of Python for functions Globus manages the function execution on any endpoint 2 2 3 Results returned to the user 26
into search index 1 Globally accessible multi-tenant service • Metadata store with fine grained visibility controls • Schema agnostic, with dynamic schema • Federated authentication integration • Query and discovery API with facets Index Index 2 Globus manages the metadata & access to fields 3 Users can query and find data of interest 29
the required steps 1 Globally accessible multi-tenant service Compute Facility v • Managed reliable task orchestration • Declarative language for flow definition • Event driven execution model • Extensible to integrate external services External Services 2 Flow run triggered by an event On-prem or public cloud storage Instrument Facility Globus reliably manages the orchestration 3 31
developed in collaboration with our Research Computing Colleagues and maintained in a library of flows allows high-speed computing to be available to a larger number of potential users. In my case, the Globus flow structure will allow me to incorporate collaborators and volunteers more easily into my research, which increases community impact and engagement. -Dan Ardia, Charles A. Dana Professor of Biology, F&M College
the Cloud (use case) Editors’ Choice: Best Use of HPC in the Physical Science Editors’ Choice: Best HPC Response to a Societal Plight Readers’ Choice: Best HPC Collaboration Editor’s Choice: Top HPC-Enabled Scientific Achievement
Delivering storage system insights • Integrated solutions using the platform services • Supporting agentic AI systems use of research CI • Meeting requirements of additional compliance regimes
research usage – Subscription required if collaborating with a commercial entity • Subscriptions enable – Enhanced features for users, administrators and developers – Removes/increase limits – Priority support • Subscription required to enable compliance regimes and access to connectors globus.org/subscriptions
• Pricing level determined by research expenditures • Separate subscription tiers for sensitive data management • Premium uplift for commercial subscribers
US national laboratories • Supercomputing facilities (US and abroad) • US agencies (and some national institutions) • Genome sequencing centers, research hospitals • Independent research institutes • Commercial research (pharma, biotech, oil & gas)