Searching for sense in the NAS market
I know I’m meant to be the voice of authority and all that, but I really can’t make any sense of the NAS market at all just now. I suspect my problem is that I’m looking for rational behaviour – just occasionally – from manufacturers and the buying public.
I suspect my problem is that I’m looking for rational behaviour – just occasionally – from manufacturers and the buying public
These last few weeks have been very eventful, because they saw the release of several NAS boxes based (finally) on Intel’s Atom platform. PC Pro has been reviewing NAS devices for a while, both in our Enterprise section and the main reviews section of the magazine, and we’ve divvied them up between sections on the assumption that readers will know what distinguishes big from small and what these different market sectors require from such devices.
I haven’t been making much noise about NAS in this column, mainly because most of my clients’ data processing requirements are multilayered, the very opposite of NAS: they need their servers to earn their keep by being able to do several different jobs at once.
It’s my basic and inescapable observation that most of these clients’ servers will struggle to average 10% processor utilisation over their working day, never mind 30%, which would be quite exceptional.
If you find a pair of idle hands in most small-to-medium businesses then it isn’t long before they’ll be found a job to do – that’s how business works and especially how network-savvy IT management works, the mark of a well-placed technology investment.
It’s easy to increase the utilisation of a general-purpose server above that low baseline by consolidating jobs on the under-used box, but you can’t do that with NAS devices. They have one job to do – in fact, the most common use I find for NAS boxes in actual businesses (as opposed to their IT managers’ home entertainment systems) is as single-session backup devices for their “real server”.
Now that’s a bit of jargon that demands unpacking to make its point. When I say “backup device”, I don’t mean something that’s capable of taking over all the jobs the server it fronts for is able to do.
I often worry that the less clued-up buyers imagine – on the basis of some article they glanced at in WHSmith while waiting for their train to be announced – that this is really what “backup device” means. If you know someone who believes this, please don’t laugh at them but put them straight in a hurry.
In the most realistic usage model a NAS is called a “backup device”, because it can be used as the target for a disk-imaging utility such as Acronis, Paragon Drive Backup or your own personal favourite.
At a couple of my clients’ sites, they’ve all been taught how to use the Acronis bootable CD to back up the server at the end of each working day. This works well as a server backup since Acronis can either reboot or shut down the machine it’s working on once the backup finishes.
Plus, when there’s only one machine pushing backups at the NAS box, that too will perform well enough (close to the maximum speed for the LAN/switch combination these guys were running in fact).
That’s the other special jargon that needs clarification – the phrase “single-use backup device” actually means “don’t try to hit a small-beer NAS box with multiple transactions”. It’s this second condition that leads to so many firms possessing a thin, rack-mount box shut down with all its lights off, about which their IT team will sadly shake their heads and say: “ah yes, we never could make that work right”.