Recently companies like
have joined the slow moving ATA disk network storage bandwaggon. This is the
one in which which you can buy a SAN or NAS stuffed with low cost ATA connected
disks, instead of faster and more expensive SCSI or fibre-channel models.
The sales pitch goes something like this. The basic low cost disk drives which
you find in a new PC give adequate performance running at 7,200 RPM. Although
you can buy disks which run more than twice as fast - the cost penalty is not
worthwhile in many real-life applications.
If the accountant is the
person in charge of buying storage in your organisation, then this argument put
forward by EMC says it all. The ATA populated CLARiiON can slash "50% of
the list price in large systems configurations". ...And a 50% reduction off
an EMC price is not an inconsiderable sum.
This is not a new idea.
Nexsan Technologies was
the first company to put its head over the parapet a couple of years ago. The
idea that anyone might knowingly want to buy a disk storage system which
was not the absolutely fastest was novel at the time.
Surely, when you
buy a new computer system you want it to be as fast as you can afford. This is a
classic buyer argument. After all, you have to live with the system for many
years, during which demand will increase and technology will pass you by. Not
necessarily so - said Nexsan. There are lots of situations when the disk
storage doesn't have to be so fast - such as archiving data onto disk, instead
of tape for example. That was still a theoretical consideration to users a few
years ago, because no one was doing this very much.
But the germ of
this idea spread and took root. Last year some bigger manufacturers started
doing pretty much the same thing.
their product BladeStore.
called theirs NearStore, while
ASACA went one step
further with its 48TB FireFly by using Serial ATA hard drives, which can be
switched off to save power until the data is actually needed.
all that, the ATA network storage market has been slow to take off, and I think
that owes more to emotional factors than the drag from a sluggish economy.
When you buy a car, you know that most of the time it will be parked
and doing zero miles per hour. When you're actually driving, the combination of
traffic congestion and speed limits, means that you may only average 20 MPH in
the town and maybe no better than 50 MPH on the open road. But most drivers
still feel more attracted to a Ferrari than a Skoda.
I'm a cautious
(read - slow) driver. And I'm very happy with my Renault Clio which my wife
mockingly refers to as being powered by a rubber band. It is true, that recently
I borrowed my wife's Audi A6, in which the engine alone cost more than my entire
car, and found I could comfortably knock 20 minutes off the 2 hour journey time
to see my sister, even while keeping to the speed limits and sticking with the
winding cross country route which I prefer. That's because, with the Audi, I
could overtake slow moving tractors or horse cars. That's something which the
rubber band powered Clio can't safely do.
And maybe that kind of worry
is a factor slowing down the new ATA network storage market.
The end to
end performance of most systems is not driven by the average speed, but by
peaks and bottlenecks.
Even if the new low cost ATA disk based
network storage systems have been purchased for one application, the fact is
organizations change, and there's the suspicion that the accountant will say -
That's it! No more IT budget this year. Out of necessity you could end up
redeploying it your main online storage. It may be faster than what you have in
place today, and even if it does what is claimed when doing an incremental
backup there will still be the niggling doubt that it will get you up that hill
fast enough when you really put your foot down in an emergency.
take more than numerical arguments to lift this market off to the next level.
That may only happen when there are enough pioneers out there who have the
experience of using them and can share their experiences. Until that time users
are no more likely to buy storage systems based on gigabytes per dollar, than
they buy cars based on miles per gallon fuel efficiency. Emotion, not logic,
sometimes gives the right answers.
Disk to disk backup