The Shared File Server is Dead?

Is it time that the file server concept was finally killed off? Is the continued use of this most ancient of network concepts holding us all back? Are they so embedded that we shall never get rid of them?

Some assumptions taken early in the technology cycle continue to influence the current technology even though they are no longer relevant. There is a well told story which links the width of a horse through to size of a train. This is apparently only half-true but serves my purposes well here because it shows that once a standard is set it is built upon, and on, and on, and may never ever be changed. So it is with the file server.

I’m not suggesting that things need to be changed just because they have been around for a while, thought I could construct an argument to that effect. In the normal course of life though something happens which changes the basis for a standard in a disruptive way and the old standard is simply left behind. In the UK the cart used to be the way to move things around, and then along came the canal, but before the canal was even fully established along came the train. The old standard may still be around, but it is no longer the standard.

The shared file server concept has been alive and well since the creation of the first network. In simplistic terms, if you wanted to share a file with someone you put it on a file server so you can both get access to it. The first and simplest form of IT collaboration, even pre-dating email. We have moved between different variant but in essence they all allow us to store files and get them back again.

This simple centralisation concept has allowed us to build a whole set of working process and technologies that have little to do with file services but build on its bedrock. We ask people to store data on the file server so we don’t need to worry about backup on the clients. We ask people to store data centrally so we can protect it with RAID systems and other resilience techniques. We ask people to store their data on the file server so we can rebuild their client device without worrying about data.

At the same time we have realised and lived with the problems of data centralisation. Moving all of the data to the centre has meant that all of the growth is centralised too leaving many organisations with an impossible task trying to keep up with the growth. Centralisation of the data has established massively complex security structures making restructuring of the data practically impossible Centralising the data has moved, to a significant extent, ownership of the data to the centre to; but because the centre doesn’t actually own the data it doesn’t know what to do with it making policy definition impossible. Centralisation of the data has also enforced a single view of the data structure which serves the needs of only some members of the team who are using the data (some people think in applications rather than in process for instance).

People have come up with all sorts of ideas to resolve these problems. Most of these ideas have involved a movement of the data out of the file systems and into some application or database. While working in some niches they have done little to curtail the continued growth of the file server. has Oracle iFiles (as it used to be called) actually removed a significant percentage of the file servers out there, how about Documentum, or any other Document Management system. Even people’s reliance upon email hasn’t slowed the consumption of the file server all that happens is that people store the data in the email system and on the file service.

So why do I say that the file server is dead? What is the disruption that is coming that will remove our reliance upon it? What technology can this idiot be talking about?

Well it’s not a technology.

It’s not some new software.

It’s society – yes society.

Or more specifically, the changes in the modern work environment leading to more pervasive home and remote working.

I regard myself as a bit of a Edge Case when it comes to home working, but I’m only one of several million. I used to architect and deploy file servers by the bucket load; since working at home I haven’t touched a single one. The file service has changed from being a service to being a constraint. In order to use the file service I need to embed myself and my device wholly into the corporate IT from the outside and it’s just too difficult. If I want to use a file service I need to start a VPN into my corporate environment. If I want to access the Document Management system I don’t, if I want to use any of the LOB applications I don’t, if I want to use email I don’t. So why should I bother? What used to be the easiest place to store data has become one of the most difficult. Perhaps there are good reason for me to use the shared file services available to me, but I don’t. It’s easier for me to share data into the places that I should have been using all along, so I have changed a long established process and started using the Document Management system.

A couple of years ago I would have told you that you were bonkers if you told me that I would be writing this post.

Where I have come many will follow.