Share
Data Unchained
Making the Potential of pNFS a Reality! w/ Trond Myklebust
Ep. 24
•
We are excited to welcome Linux NFS Kernel Maintainer and CTO of Hammerspace, Trond Myklebust, to join us on this episode of the podcast! Trond and Molly discuss the Linux community and Trond's journey from aspiring Particle Physics Ph. D. to Linux Maintainer and innovative industry technology visionary.
Trond has dedicated a career to building open source software and driving innovation in data technologies. In this episode we discuss the evolution of high-performance file systems. Historically, parallel file systems have required additional client software to be loaded on each client machine that needs to work with high-performance data sets. Added software can be difficult to get approved by security team standards, slow to be added to workstation images, and is typically charged per client software instance. These challenges have made it difficult to give access to all users and applications that could derive value from the data sets housed in the parallel file system.
All of this began to change with the vision of the Linux community developing an embedded parallel file system as part of the NFS protocol. With the creation of pNFS (parallel NFS), compute-standard Linux clients now can read and write directly to the storage, and scale performance linearly and near infinitely. Expensive and proprietary software is no longer needed to create a parallel file system. pNFS is built into open standards.
All these topics and more as we dive deeper into the data driven world on this podcast episode of Data Unchained!
#data #pNFS #NFS #Linux #Maintainer #Community #decentralizeddata #datastorage #storage
Cyberpunk by jiglr | https://soundcloud.com/jiglrmusic
Music promoted by https://www.free-stock-music.com
Creative Commons Attribution 3.0 Unported License
https://creativecommons.org/licenses/by/3.0/deed.en_US
More episodes
View all episodes
116. Is SaaS Obsolete? The Future Of Data, Local Compute, & AI w/ Bob Heckel
30:58||Ep. 116Enterprise AI is forcing organizations to reconsider how much of their technology, data and infrastructure they should own. Bob Heckel, CEO of Computer Visionaries AI, joins Molly Presley on Data Unchained to explain why he believes traditional SaaS is losing its place in an AI-native business environment. They discuss how computer vision and edge inference can improve retail, manufacturing, logistics and workforce operations, as well as why real-time processing often belongs closer to where data is created. Bob also explains how open-source models and local infrastructure can reduce token costs, address GPU constraints and give companies greater control over their data.Learn more about Computer Visionaries AI: https://cvaius.com/Connect with Bob Heckel on LinkedIn: https://www.linkedin.com/in/heckel/Contact Computer Visionaries AI: sales@cvaius.comLearn more about Hammerspace: https://hammerspace.com/
115. How AI Data Can Use Network Telemetry Without Sacrificing Reliability & Control w/ John Capobianco
46:19||Ep. 115Network infrastructure generates enormous volumes of data, but much of it remains difficult to access, interpret, and use. In this episode of Data Unchained, John Capobianco, Head of AI and Developer Relations at Itential, joins Molly Presley to discuss how AI agents can turn network telemetry into practical insights, predict equipment failures, support troubleshooting, and help organizations build reliable sources of truth.John explains why read-only agents offer a lower-risk starting point for agentic operations and how natural language could become the primary interface between people, controllers, and infrastructure. The conversation also covers GPU-cluster networking, data retention, open-weight models, RAG, Model Context Protocol, data sovereignty, deterministic workflows, and the governance required before agents are trusted to make changes.Join the VibeOps community: https://vibeopsforum.com/join/Connect with John: https://www.linkedin.com/in/john-capobianco-644a1515/Watch John’s videos: https://www.youtube.com/@johncapobianco2527Learn more about Itential: https://www.itential.com/Learn more about Hammerspace: https://hammerspace.com/
114. From Compute to Data: The Next Phase of AI Infrastructure w/ Wendell Wenjen
41:19||Ep. 114Enterprise AI is moving from experimentation to production, forcing organizations to reconsider how they store, access and manage their data. Wendell Wenjen, Senior Director of Storage Market Development at Supermicro, joins Molly Presley on Data Unchained to discuss the infrastructure demands behind enterprise AI. They examine the shift from training to inference, the growing role of storage in AI budgets, KV cache, on-premises AI and the challenge of preparing distributed enterprise data for production workloads. Wendell also explains why many AI proofs of concept never reach production, how power demands could influence infrastructure decisions and what businesses must address before deploying AI at scale.Learn more and register for the Supermicro Open Storage Summit, airing August 11 through September 3, 2026: https://thecube.net/Learn more about Hammerspace: https://hammerspace.com/
113. Is the Data Migration Nightmare Finally Over? w/ Pradnesh Patil
40:12||Ep. 113What is really slowing down enterprise AI? According to Pradnesh Patil, Founder and CEO of Altimate AI, the answer is not always the model. It is the data work behind it. In this episode of Data Unchained, host Molly Presley talks with Pradnesh Patil about the pressure facing modern data teams as AI adoption accelerates. They discuss why data engineers are buried in pipeline backlogs, how token costs can spiral out of control, and why enterprises need more open, vendor-agnostic ways to manage AI-driven data work. Pradnesh explains how AI agents can help with data pipeline creation, infrastructure optimization, governance, context management, and the painful migration projects most teams try to avoid. He also breaks down why walled-garden AI tools can limit flexibility, why context graphs matter, and how enterprises can reduce cost without sacrificing control.https://www.altimate.ai/Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US
112. Is Science Next in the AI Disruption Cycle? w/ Kevin Moore
40:45||Ep. 112In this episode of Data Unchained, Molly Presley talks with Kevin Moore, CEO of Quilt Data, about what it takes to make scientific and enterprise data ready for AI, collaboration, and discovery at scale. Kevin explains how life sciences teams are managing massive multimodal data sets across AWS, S3, sequencing workflows, lab notebooks, metadata systems, governance requirements, and regulatory demands. The conversation explores how bench scientists, computational scientists, data engineers, and IT teams can work from the same versioned, contextualized data sets without losing control of access, compliance, or intellectual property. Kevin also shares why metadata, object storage, version control, and flexible data set constructs are becoming essential as organizations move from isolated data silos toward human and AI-assisted research workflows.Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US
111. AI Has Data Problem Supercomputers Already Solved w/ Mike Heroux
39:24||Ep. 111In this episode of Data Unchained, Molly Presley talks with Mike Heroux, Senior Research Scientist at ParaTools and Scientist in Residence at Saint John’s University, about what the future of AI can learn from decades of high performance computing. Mike shares lessons from open source scientific software, the Exascale Computing Project, national lab collaboration, world models, AI for science, and the growing connection between modeling, simulation, and modern AI infrastructure. The conversation also looks at why data, software, and compute must work together, how open source became essential to scientific progress, and where quantum computing may fit into the next generation of research and engineering.Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US
110. Wrong Data In, Disaster Out: The Data Problem That Could Wreck Your AI Strategy w/ Mike Seashols & Floyd Christofferson
49:03||Ep. 110Molly Presley sits down with Mike Seashols, CEO of Secuvy, and Floyd Christofferson of Hammerspace to discuss one of the biggest enterprise AI challenges today: governance. This episode explores why enterprise AI is really a data problem, how the wrong data creates risk, and what it takes to make live enterprise data AI-ready without massive migrations or added complexity.Demo: https://hammerspace.com/hammerspace-aidp-demo/Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US
109. The Future of AI Will Belong to Companies That Make Data AI-Ready Now w/ Jack Hogan
45:31||Ep. 109In this episode of Data Unchained, Molly Presley talks with Jack Hogan of SHI about why enterprise AI success depends on more than just powerful models or infrastructure. Their conversation explores what it really takes to make data usable for AI, from governance and security to discoverability, context, and operational readiness. They also dig into why so many enterprise AI projects stall, what business leaders often get wrong, and how organizations can better prepare for long-term success. If you are trying to understand the real barriers between AI ambition and business value, this episode offers a practical and timely look at the challenge.Find out more at: https://www.shi.com/Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US
108. Global Namespace for Precision Medicine: The Data Breakthrough w/ William Baird
12:01||Ep. 108Global namespace is becoming a critical strategy for cancer genomics, precision medicine, and life sciences data infrastructure. In this episode of Data Unchained, Molly Presley speaks with William Baird from Guardant Health about managing 160PB of genomic data across HPC, on prem systems, and multi-cloud environments. They break down why 80% of this data is still active, how researchers need faster access without worrying about where data lives, and why global namespace could become a major solution for decentralized data at scale. The conversation also explores vendor flexibility, continuity, standards development, and what it takes to support life sciences workflows when the stakes are incredibly high.Cyberpunk by jiglr | https://soundcloud.com/jiglrmusicMusic promoted by https://www.free-stock-music.comCreative Commons Attribution 3.0 Unported Licensehttps://creativecommons.org/licenses/by/3.0/deed.en_US