Primary Data contributes to NFS 4.2

Primary Data (www.primarydata.com), a leader in unified parallel storage based on pNFS, continues its market penetration effort with its recent participation to the #21 IT Press Tour and the 3rd participation for the company. Great meeting with 2 key leaders present: Lance Smith, CEO, and David Flynn, CTO & co-founder.
The company has made great progress since our last visit and especially with the GA of DataSphere available since VMworld in August.

To summarize Primary Data solution - please refer to past posts I wrote in Aug. 2013, Nov. 2014 and May 2016 - let's use a few keywords: DataSphere is a software that offer a data virtualization access layer across a variety of storage devices supporting block, file and objects protocols implementing a asymmetric model aka out-of-band architecture leveraging pNFS to provide the intelligence and the parallel data transfer capability. In other words, DataSphere helps IT to protect storage investment, create data placement policy to align the value of the data with the platform characteristics and also use the data where they reside without any need for special data movement. The product provides a single global dataspace set by a set of Directors - where meta-data, policies.... reside - acting as the control plane connected to the physical storage - where data live - representing the data plane. Beyond the beauty of pNFS, Primary Data adds 3 keys elements: Smart Objectives, Data Mobility and Data Analytics leveraging all meta-data controlled by the Director. In addition, the company has developed DSX, for Extended Data Services, to provide a client data portal, a data mover and a cloud connector. Both products, DataSphere and DSX are sold via an annual subscription.

With the announcement of Primary Data's NFS 4.2 contributions - cluster server deployments support, cross back-end support and Linux integration especially Flex File in 7.3 RHEL - the company is now seen as an official leader in Enterprise Data Virtualization. Flex File is a new pNFS layout that allows aggregation of standalone NFS servers, exporting of clustered file system and per-file striping. A few other key things are added in this 4.2 release such Server-Side Clone and Copy, Application Input/Output for I/O behavior expectation (IO_ADVISE), Sparse Files ("hole punching"), Space Reservation, Application Data Block or Application Data Holes aka Block on File, Labeled NFS and Layout Enhancements. An interesting feature illustrated on the white board by David Flynn was the client local storage capability.
It was really a very good session that demonstrated if needed that Primary Data is now fully equipped to shake the storage industry palm in 2017 with a very comprehensive solution in all 3 dimension: Data Virtualization, Orchestration and Automation.
Primary Data is the perfect illustration of an Universal Data Storage approach with any to any combination of vendors, storage devices or protocols which helps IT decision makers do the right choice and protect their investment. Let's meet again in 2017 with the IT Press Tour crew.