From my POV traditional NAS is dead (and I mean dead as per mainframes or tapes… so, don’t get me wrong about this post!).
At the same time NAS is alive and kicking, it’s literally evolving into something new and will continue to be part of all Data Centers.
End user needs are dramatically changing but, as always, processes and applications that are already in place cannot be modified. This is leading to a new generation of solutions that can be accessed with legacy protocols, like a “normal NAS”, but with far more features and with the capability to solve problems which are totally different from the past.
Why I say traditional NAS is dead
The classic (and with classic I mean old) NAS architecture (a traditional box with disks at the backend and NFS/SMB at the from end) is no longer sufficient to cope with the needs of today’s organizations. That kind of Network Attached Storage is commodity now: NFS, SMB, other protocols, AD integration, data footprint reduction of some sort, snapshots, replication and so on, are implemented by all vendors. Differences between vendors are minimal and so is the perceived value. The most important feature is price and $1/GB is more than the majority of end users are willing to pay, (average prices I’m registering in the field for NAS boxes, produced by primary vendors, in the range of 100/300TB of usable space).
Not only NAS
Modern organizations want more from storage systems and the meaning of the term NAS can actually differ according to the context of where it’s being used (BTW, this is also happening to other kinds of storage systems). User choices are driven by their needs and the term NAS alone is too generic to cover all those vertical needs.
Do you need plenty of local space and performance? scale-out NAS is perfect for you.
Do you need VMware/Oracle/Hyper-V/Whatever performance and integration? integrated NAS (hybrid backend and sophisticated features) is what you are looking for.
Do you need a tailored solution? open source ZFS based NAS is excellent.
Are you thinking about hyper-convergence? most of the products are NAS VSAs.
Do you need to manage several remote offices? NAS gateways with object storage at the backend is a solution.
Big data? Metadata? Active archive? Primary fast storage? Only a backup repository? you name it. Many needs can be delivered with a NAS but not a traditional NAS? At times it’s not even referred to as NAS anymore.
Sometimes you need a legacy access method to allow users to work as they always have, other times you need it only to provide a seamless upgrade path to new technologies (think about a data-ingestor for an object storage for example).
NAS, especially traditional NAS, is not adequate to deal with devices like smartphones and tablets, but all environments are mixed now and you need solutions to provide the best user experience on the various platforms.
It’s no secret that I’m an Object Storage fanboy and I’m convinced that NAS is only becoming a part (an access layer) of a bigger, broader storage infrastructure. For example, in most of the cases, a sync&share service needs to access the same data which is accessed from a NAS… or you need to provide an abstraction layer capable of treating all devices and PCs in the same way (currently I’m looking at Maginatics and I’m finding its approach brilliant, to say the least!).
Why it matters
NAS is not dead but in this “software-defined” world it needs to be rethought from the ground up. I’m firmly convinced that most of the organizations have to think differently about their storage strategy for the future: NAS, SAN, unified storage and other traditional concepts simply cannot keep up with the new needs… and new paradigms are emerging.