Skip to content
Artwork for Techsplainers by IBM
Techsplainers by IBM · Friday · 9 min

What is NVMe?

This episode of Techsplainers explores NVMe, or Non-Volatile Memory Express, the protocol designed to help systems get more performance out of modern non-volatile storage like flash memory and SSDs. As the final episode in this data storage series, it connects the dots between storage media, system architecture and real-world performance. The episode explains why older storage protocols were built for mechanical hard drives and how that created bottlenecks once flash storage became much faster. It then breaks down how NVMe uses the PCIe bus, reduces overhead and supports far more parallel commands than earlier approaches, helping lower latency and improve throughput for demanding workloads. Listeners will hear where NVMe matters most, from enterprise databases and AI workloads to everyday laptops and desktops, and how it differs from NAND flash itself. The discussion also introduces NVMe over Fabrics, which extends NVMe principles into networked storage environments. By closing out the storage series with NVMe, the episode highlights a central takeaway from the broader run: storage performance depends on layers working together, from architecture and protection to media and protocol. Learn more in https://www.ibm.com/think/topics/nvme Find more information at https://www.ibm.com/think/podcasts/techsplainers Narrated by Elly Trickett "AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication."

0:00-9:52

transcript

No transcript — this publisher did not publish one.

show notes

This episode of Techsplainers explores NVMe, or Non-Volatile Memory Express, the protocol designed to help systems get more performance out of modern non-volatile storage like flash memory and SSDs. As the final episode in this data storage series, it connects the dots between storage media, system architecture and real-world performance. The episode explains why older storage protocols were built for mechanical hard drives and how that created bottlenecks once flash storage became much faster. It then breaks down how NVMe uses the PCIe bus, reduces overhead and supports far more parallel commands than earlier approaches, helping lower latency and improve throughput for demanding workloads. Listeners will hear where NVMe matters most, from enterprise databases and AI workloads to everyday laptops and desktops, and how it differs from NAND flash itself. The discussion also introduces NVMe over Fabrics, which extends NVMe principles into networked storage environments. By closing out the storage series with NVMe, the episode highlights a central takeaway from the broader run: storage performance depends on layers working together, from architecture and protection to media and protocol. Learn more in https://www.ibm.com/think/topics/nvme Find more information at https://www.ibm.com/think/podcasts/techsplainers Narrated by Elly Trickett

"AI tools may be used to transcribe this episode and support selected stages of the production process. All AI-assisted content is reviewed by the production team before publication."
links2

more episodes

All episodes