Vendor lineage

NetApp

Named itself after exactly what it sold, which turned out to be most of the strategy.

Network Appliance was founded in Sunnyvale in 1992 by David Hitz, James Lau and Michael Malcolm, on the observation that building file storage out of general-purpose servers was harder than it needed to be. The product was an appliance: a box that did one job, with an operating system - Data ONTAP - written for that job alone.

The name said what it was, which was an advantage in a market where everything else was assembled from parts. The first filer shipped in 1993, Sequoia funded the company in 1994, and it went public in 1995.

Its main competitor at the outset was Auspex Systems, where both Hitz and Malcolm had previously worked - which is the ordinary way storage companies begin, with people leaving one to build the thing it would not.

The dot-com years took NetApp past $1B in annual revenue. When the bubble burst, revenue fell to around $800M in fiscal 2002, and the recovery from there was slow and unglamorous rather than dramatic. It changed its legal name from Network Appliance to NetApp in 2008.

Two threads worth following from here. Michael Malcolm left the chief executive role in 1994 and founded CacheFlow in 1996, which became Blue Coat Systems - and in 2006 NetApp sold its NetCache product line to Blue Coat, so a company sold a business to the firm its own co-founder had built. Malcolm's earlier career had run through Sun Microsystems and Tandem Computers, both of which appear on this timeline having ended inside somebody else.

Hitz and Lau shared the IEEE Reynold B. Johnson Information Storage Systems Award in 2007. Hitz retired in 2019.

Founding stories

1992

Network Appliance

Sunnyvale, California · Founders: David Hitz, James Lau, Michael Malcolm

Incorporated in April on the proposition that file storage should be a box that does one job rather than a general-purpose server persuaded into doing it. The stated ambition was to make storage as simple as an appliance, and the company took its name from the category it intended to create.

The timeline

Founded
In Sunnyvale, by three people, two of whom had come from Auspex.
The first filer
A dedicated network file server, sold as a finished thing rather than as parts to assemble.
Sequoia funds it
And Michael Malcolm leaves the chief executive role the same year.
Public
Three years from founding, on a product line one deep.
The correction
Fiscal 2002 revenue around $800M, down from above $1B.
Network Appliance becomes NetApp
The legal name changed.
The pivot showing in the numbers
Quarterly revenue of $1.64B, all-flash on an annualised run rate of $3.8B, and cloud storage services growing more than 40% year over year - the software business outgrowing the hardware one, on a company whose founding thesis was the hardware.

Flagship products and solutions

ONTAPThe operating system, and the actual product. It has three parts: the execution environment, the file system, and the data protection layer - and it now runs on NetApp hardware, on commodity servers, and inside all three major clouds.
AFF and FASThe all-flash and hybrid arrays. The A-series takes the mission-critical workloads, the C-series the general-purpose ones, both speaking file, block and object from the same system.
StorageGRIDObject storage for the data that is kept rather than used - archives, compliance, the material nobody reads until somebody has to.
Cloud Volumes ONTAPThe same operating system running inside AWS, Azure and Google Cloud, so a workload can move without changing how its storage behaves. This is the pivot, expressed as a product.
FabricPoolAutomatic tiering of cold blocks out to object storage, which is the unglamorous economics of keeping expensive media for data that is actually being read.
KeystoneStorage sold as a subscription rather than as capital equipment - the same shift every infrastructure vendor has been pushed into by the cloud's pricing model.

Key innovations

Never overwrite a blockWAFL writes changed data to a new block and moves the pointer, rather than rewriting the original. That single decision is the foundation: it makes writes fast because they can go anywhere convenient, and it makes crash recovery quick because the previous consistent state is still intact on disk.
Snapshots as pointer arithmeticBecause nothing is overwritten, a point-in-time copy is a copy of the pointers rather than of the data. That makes snapshots effectively instant and nearly free in space - and it is why a feature every competitor eventually had to answer was, here, a consequence of the file system rather than a product bolted onto it.
Metadata kept in ordinary filesMost file systems fix their metadata structures into the on-disk format, which makes changing them a migration. WAFL holds metadata in regular files, so the format can evolve without moving anybody's data. That is a large part of why a design from 1992 is still shipping.
Selling the software once the box stopped being the pointA company named after an appliance now sells its operating system to run inside other people's clouds, and that line is growing faster than the hardware. It worked because the value was always the file system rather than the sheet metal around it - which was not obvious at the time and is the reason the transition was survivable.

Main markets

Enterprise storage across every sector that keeps more data than it can move: financial services, healthcare, media, research computing, and increasingly the data pipelines behind machine learning, where the problem is feeding accelerators fast enough rather than storing cheaply.

Reported around $6.6B of annual revenue, with all-flash and cloud services carrying the growth. Its competitors are the other large array vendors, the hyperscalers' own storage services, and the software-defined entrants that make the same argument NetApp made in 1992 against the incumbents of that decade.

Analyst standing

Placed among the top three global enterprise storage vendors in IDC's quarterly tracker for the third quarter of 2025, and named a Leader and Fast Mover in GigaOm's primary storage assessment.
The assessment worth noting is not the ranking but the shape: a thirty-year-old hardware company whose fastest-growing line is software running on infrastructure it does not own. Most of its contemporaries from 1992 did not manage that transition, and several of them appear elsewhere on this timeline having ended inside somebody else.

Acquisitions

2003 Spinnaker Networks $300M
Clustered storage technology.
The clustered Data ONTAP architecture.
2008 Onaro $120M
Storage service management.
SANscreen, later OnCommand Insight.

NetApp ↗

From the company

Sources