Customers have the power in any relationship with vendors these days,
and that's a good thing. One example of how much things have changed on
this front is iSCSI.
Back in the 1980s, an often-used expression was: "Nobody ever got fired
for buying IBM." That was the explanation from some customers about why
they bought Big Blue. It was a "safe" decision, in large part because
everyone in top management had heard of that vendor and it was a choice
not usually met by any resistance. IT-ers got their purchase requests
approved and everyone (especially IBM) was happy.
Nowadays, though, that type of blind allegiance to a vendor or
particular set of products is simply not the norm. Sure, it still exists
in some remote outposts but it's by far the exception and not the rule.
Nor will most people invest substantially in any technology simply
because a vendor - even a big one - is pushing it. There's got to be a
business driver.
Which brings us to iSCSI, the new standard for using the Internet as a
mechanism to move information to and from various storage devices.
Microsoft is the latest major vendor to jump on the bandwagon. It put
iSCSI drivers into its Windows Server 2003 family of products, and
recently announced that products from more than a dozen storage
networking vendors have been certified as working with Microsoft's iSCSI
Software Initiator.
Don't get me wrong. Any time a group of suppliers agrees on a tech
standard, that's usually good for customers. Standardization leads to
more product choices, ultimately lower prices and is evidence of
capitalism working well. But that doesn't necessarily mean customers
should buy into any of this - at least not yet.
Eventually, iSCSI will give low- and mid-range NAS a good run,
especially in medium-size businesses that have multiple departments or
offices that need to share storage. The iSCSI standard will go a long
way toward eliminating the "islands of storage" problem so pervasive
with most NAS devices. And it will rival NAS, and low-end SANs, in terms
of ease of installation and low cost and, ultimately, performance.
But savvy customers know it isn't there yet. A reader emailed me a few
months back, asking for production implementation information about
iSCSI. I did some searching and came up empty, convinced that there
wasn't more hands-on experience being reported because there weren't
that many users with that level of experience quite yet.
It's a new standard, and as such it will be proven out in the large
shops with technology R&D teams that have the time and budget to invest
in new ideas. It will take a while for these success stories to trickle
down to the real power users for iSCSI, the medium-sized enterprises. As
that occurs, iSCSI will be implemented in the time-honored way that most
new tech is: one project at a time, with specific business requirements
and even more specific payback periods.
Customers will sort this out, and it will take time. No matter who's
waving what flag.
