F5: The last vendor standing in file virtualisation?

Since 2005 file virtualisation has gone from emerging technology to a field with one active vendor. What has caused the demise of file virtualisation and why does F5 persist with it?

The demise of AutoVirt in January 2012 thinned out the number of vendors selling discrete file virtualisation products to just one.

It wasn’t always that way. File virtualisation was once a rising storage market category, and in the middle of the last decade there were numerous file virtualisation product vendors in existence.

The basic premise of file virtualisation -- although different vendors did things differently, of course -- was to insert a layer between clients and the storage, creating a storage pool from among heterogeneous storage systems. By introducing this layer between client and storage, users were able to manage files from a central location rather than having to leap between servers and/or storage devices.

In file virtualisation, disparate devices hold the data, but a global namespace indexes files and enables storage consolidation. Files are not tied to specific hardware and are very much easier to find. Migration of data is also made simpler, and file virtualisation can even remove the issue of duplicate file names, as they are indexed by the virtualisation layer under a unique identifier.

When file virtualisation was unveiled to the market in the mid-2000s, there were numerous players, such as NeoPath Networks, Brocade and Attune, selling their wares. So, what’s happened to the market?

Some products disappeared while the wave of hype was at its peak. NeoPath, for example, shut down in 2007 not long after it was bought out by Cisco. Brocade discontinued its file virtualisation products in 2009, and Attune was swallowed up by F5 in the same year.

We can find only two companies still giving life to this near-extinct technology.

F5 appears to be the last vendor actively marketing file virtualisation. The vendor has its ARX series of appliances that work with Windows, Unix and Linux platforms, and when it acquired these from Acopia to add to its own portfolio, it even touted the “death of the NAS” as part of its marketing campaign.

EMC has its Rainfinity appliance, but it has been refocused as a migration tool add-on to services, said Simon Robinson, a senior analyst at The 451 Group.

“I would say the only one who really makes a big deal out of it anymore is F5, but even for them it's not a huge business.”

Nigel Burmeister, F5 director of product marketing, said file virtualisation is still a key part of the company's portfolio and claimed that F5’s technology has survived by virtue of its uniqueness.

“It’s our in-band technology that allows it to be a real-time device,” said Burmeister. “All other approaches [from our rivals] were dual- or out-of-band, and that limits what you are able to do. It limits you to migration use cases and also has severe limitations on things like scale. If you are in-band, you have the scale and form factor that allows you to address a broad cross section of accounts.”

However, the fact that so many other companies involved in file virtualisation have gone under cannot be ignored. Independent technology analyst Claus Egge puts it down to other storage technologies solving the issues file virtualisation attempted to address without the need for the extra layer it involves.

The likes of scale-out NAS and object-based storage provide the same global namespace, for example, but give better capabilities for larger workloads being processed in one go.

“[These technologies] basically solve issues that the traditional file systems struggle with, and they replace the existing storage, eliminating duplication,” said Egge.

Yet, F5’s Burmeister believes there is still a need for different delivery models of file virtualisation, rather than it just being a feature within a storage vendor’s product.

“We chose to deliver [file virtualisation] technology as a standalone appliance because it brings customers a lot more flexibility when it comes to storage choices,” he said.

“If you are a storage vendor looking to implement this technology, you would look to deliver it differently,” he added.

He admitted that scale-out NAS gives many of the same features as file virtualisation tools and that it made sense for the big storage vendors to include those features as part of their storage products.

“However, we feel there is another subsection that requires the flexibility to choose different storage technologies because storage changes so quickly nowadays,” he added.

Essentially, the selling point of having a separate appliance, according to F5, is that whatever mishmash of storage products you have within your data centre, the file virtualisation layer is compatible with them all and can bring all your files together.

This may be F5’s view, but the industry seems to have moved on, and adding yet another layer of virtualisation to an already busy data centre goes against what enterprises are looking for, said Egge.

“I think the proposition was not all that compelling,” he said. “Buying good storage arrays and then having them managed by a file virtualisation layer duplicates function and cost.”

So, F5 is the last vendor standing in the file virtualisation game. Let’s see how long that’s the case for.

Jennifer Scott is a technology writer.

Read more on SAN, NAS, solid state, RAID