In this week’s episode, the team talk to Rob Peglar, Senior VP and CTO at SymbolicIO. The conversation covers persistent memory and in particular the NVDIMM format of devices that provide persistent storage in a standard server DIMM slot. Rob explains how the standards are set through JEDEC, with a description of the three types – NVDIMM-N, F and P.
The discussion continues to use cases, which seem mainly to be for any application needing low latency with local server persistence. Databases and analytics are the main beneficiary, however Rob believes that the hyper-scalers may be able to take advantage of the technology in the future.
A few references are made during the podcast. The SNIA Persistent Memory sub-site can be found here at https://www.snia.org/pm. Here’s a link to the Persistent Memory Summit – https://www.snia.org/pm-summit. It’s actually Linux kernel 4.2 onwards that adds NVDIMM support as well as Windows 2016. This Microsoft webinar provides some detail on how NVDIMM can be used with Storage Spaces Direct – Using Non-volatile Memory (NVDIMM-N) as Block Storage in Windows Server 2016.
Elapsed Time: 00:28:46
Timeline
- 00:00:00 – Intro
- 00:00:54 – Persistent Memory defined
- 00:02:00 – Why use it? What is it for?
- 00:04:40 – Why not just use a honking great battery for protection?
- 00:05:20 – Standards from JEDEC
- 00:06:00 – Defining the NV-DIMM formats. N, F and P.
- 00:12:00 – Could 3D-XPoint justify its own definition?
- 00:13:35 – How are these devices supported by the hardware?
- 00:16:00 – What are the application use cases for persistent memory?
- 00:19:00 – Could I trust having NVDIMM as my only source of data?
- 00:23:00 – Who is buying and using NVDMM today?
- 00:27:30 – Wrap up
Copyright (c) 2016-2018 Storage Unpacked. No reproduction or re-use without permission.
Podcast: Play in new window | Download