Physical storage? That is just not logical

I am currently averting another impending drive space shortage, thanks to Spiceworks and its alerting system – I need to shuffle files around to create more storage until I can take the server offline to sort out the problem properly. A long time ago we might have celebrated averting this potential disaster but these days it is just not newsworthy and drifts by silently. After all, it isn’t like we are short of space – not by a long shot – we have oodles of physical storage. The problem is: physical storage does not easily translate into available storage.

We just don’t have storage where, and more importantly when, we need it.

The storage space we do have available fills quickly. Of course there is the space utilised by the user base but we also need to factor in the indirect storage usage which includes duplicate emails, attachment saving, office back up files, temporary files and documents stored in electronic form - especially with our digital scanning project in operation with renewed fervour. Another factor to add to is our requirement for mirrored storage, multiple levels of archive storage, and file version history storage. One last problem is temporary projects - as part of the ongoing email ingestion project I need to find a large amount of storage to temporarily store extracted emails whilst I organise them into a suitable format – where do you get 100GB of storage when you need it quickly?

Our drive space utilisation, across the various servers, can be broadly broken down into a number of groups:

• Auditing/Log Files
• Regular File Backup
• Mirroring
• Server Files
• User Files
• Archiving

In these days of making the best of what we have it seems somewhat strange there is an apparent lack of economical and efficient methods for allocating and managing drive space – as a centralised unit. We have a dedicated NAS giving us nearly 3TB of physical storage but this cannot be used for everything (many applications still require local storage). When we take the dozen or so breeds of servers (from Unix to Windows), all with their own drives and storage requirements, we get almost that much physical storage space again.

These drives all need managing separately and the impact of a drive failure on one device, for example the email server, can be quite different to the impact of a drive failure on the NAS for example. Once drive space is allocated it cannot be easily reallocated or increased/decreased – so we often have to over-specify our requirements - leaving a lot of stranded space around the place that could be put to better use if it could be managed centrally. The whole thing is very inconsistent.

This is a shame because drive space physically exists – we just cannot get at it easily.

Of course, we could create a vast array of shares, to cater for the stranded space, but who wants to manage that?

Certainly virtualisation has come to the rescue of low budgets and available processor power – we can squeeze every last drop out of what we have on our server processor. Can we do the same for storage space?

I am not entirely convinced I can talk the business people into buying into a solution to fix a problem when the answer is likely to be “buy more storage” - especially if we get dragged into licensing issues and limitations. The one thing that really enabled the virtualisation project was Microsoft Virtual Server – all we had to provide were server licenses.

All these physical drives are constantly spinning, playing an odd game of high risk roulette, whilst wasting power – I imagine we could save quite a bit of energy if we could find a smarter way to manage storage. Are there any smart solutions out there?