Pittsburgh, PA – September 26, 2025 (Updated 10/06/2025) — The Massive Storage Systems and Technology (MSST) conference returned to its long-time venue at Santa Clara University’s School of Engineering. This year’s edition stood out for running in September (rather than late spring) and for its focus on NFS and distributed file storage, marking the 40th anniversary of the protocol first introduced in 1985.

Speakers revisited NFS’s evolution through versions 3, 4, and pNFS, highlighting how open-standard design has kept it relevant across millions of deployments.

pNFS emerged as a central theme. After a slow commercial start, AI workloads are driving broader adoption and innovation. Presentations explored new implementations and optimizations for modern data pipelines. Beyond NFS, sessions covered the broader storage ecosystem: S3 performance for AI, scalable indexing and metadata management, persistent-memory file systems, automation in benchmarking, reliability insights, advances in HAMR technology, and TCO perspectives for large-scale capacity planning. Tape storage also remained a topic of interest, with updates on LTO-10 and remote-access capabilities.

Looking ahead, organizers indicated MSST will return before summer 2026, likely the week of June 15 (to be confirmed).

Highlights from MSST 2025

• NFS @ 40: Historic perspectives from early contributors and maintainers; renewed focus on pNFS layouts for throughput and scale.
  
  
• NFS/RDMA & stack tuning: Deep-dive sessions on NFSoRDMA, xiRAID + XFS, LOCALIO, nConnect, and multipathing.
  
  
• Object storage for AI: Focus on S3 performance and deep-archive integration strategies.
  
  
• Media roadmaps: Updates on HAMR technology; LTO-10 tape advances and remote-access options; insights on QLC and emerging form factors.
  
  

What it means

MSST 2025 highlighted how AI is reshaping file and object architectures—reviving pNFS interest, driving S3 performance innovation, and keeping long-term storage technologies (HAMR, LTO-10) central to cost efficiency and scalability.

‍