News Stay informed about the latest enterprise technology news and product updates.

Bad Packets: Recyling old ideas

Is IBM's iBoot a technological breakthrough in networking or just an old idea recycled?

E-mail Wes Simonds

The late novelist Douglas Adams (1952-2001) once wrote, in reference to technology, that many of the best ideas are simply back-to-front. That is, you (1) Take an old idea, and (2) Reverse the logic.

According to Adams, you sometimes find yourself innovating spectacularly as a result.

One of his books described a task planner which operated in this way. You told it what you wanted, and it worked backwards in logical steps until it arrived at what you already had, thus yielding the simplest method to achieve virtually anything at all.

This was clearly a significant improvement on task planners of the day, which began with what you had and worked forwards logically until, well, until they crashed. (My recollection is that Adams' digital genie was sold mysteriously to the U.S. military and subsequently used to conceal or justify covert operations of all types.)

For more information
Bad Packets: Chamber's vision sells industry short IBM's iBoot home page Best SearchNetworking Web Links on network designs

 Today's technology giants don't seem to turn out such innovative products. It's been my experience that they're a lot more likely to develop back-to-back ideas. The back-to-back idea is even simpler than the back-to-front idea. To develop a back-to-back product, you (1) Take the stuff that came out five years ago, and (2) Do not change it.

Well, actually you rename it, shrinkwrap it and sell it again. I guess I might also call it the Back to the Future principle since it's effectively commercial time travel.

Microsoft, of course, does this better than anyone I can think of. To my knowledge, Microsoft has never actually invented a major category of software, as Aldus did with PageMaker or Dan Bricklin did with VisiCalc. What it does instead is to notice what other people succeed at selling and then imitate that as quickly and overwhelmingly as it can by integrating it with Windows. On the rare occasions when it tries to innovate, as with Microsoft Bob, it generally does not do so successfully.

Last week, I saw the back-to-back principle at work again, when IBM announced the development and future anticipated release of something called iBoot.

In the company's words, iBoot "pushes iSCSI technology even further, opening the door to the exciting possibility of the diskless computer. IBoot technology allows a diskless boot of either Windows XP, Windows 2000, or the Linux operating system from an iSCSI target machine remotely located over a standard IP network."

In other words, it permits what you might call a "thin client" or "network computer" which doesn't have a "disk drive" or "software" and instead allows network administrators to manage users' "operating systems," "applications," and "data" from "servers."

Ah, me. In only my last column, I said that Larry Ellison's concept of the network computer (circa 1997, or phrased another way, five years ago) had failed because it caught fire in nobody's imagination except Ellison's.

I see now that I spoke too soon. Clearly IBM's imagination was as ignited as the day is long. (And I live in Austin, Texas, and the days are really very long in June.)

I was at least right that the network computer never really sold over the course of the last five years, and you might reasonably extrapolate that if iBoot is basically the same thing, it's not going to sell either. But why?

Well, on the evolutionary scale of billionaire CEO nutjobs who buy jet fighters and get sued for dating subordinates, Larry Ellison is almost, but not quite, in a class by himself. And when he's not talking about enterprise-class database solutions, it's very tempting to blow off whatever he says.

Many people did.

Then there were those who thought that the network computer was a doomed approach because it was simply a recreation of the mainframe paradigm beyond which everyone had already moved. That was a hasty conclusion, to my mind, because while the storage of the network computer is centralized like a mainframe, the processing is not; each terminal does its own thinking.

In the case of iBoot, I fear I see no such fundamental innovation. I see iBoot as Ellison's network computer with iSCSI and that's about it. In other words, I see iBoot as less clever still than its failed predecessor.

One more argument against the thin client was cultural. People in positions of power, went this argument, such as IS directors who could actually take such an idea and make it a viable reality, ignored it specifically to maintain their power.

If half your staff and half your budget is given over to client-side problem-solving, and you deploy network computers that never break, and thus eliminate client-side problems. What happens to your staff and budget? Are they cut in half?

Gee, that's no fun.

Microsoft saw things in simpler terms even than this. Bill Gates said people liked traditional PCs, wanted to keep using traditional PCs, and in fact would keep using traditional PCs, and that was that.

Which argument applies to iBoot? I say all of them. It's certainly never going to sell.

However, I have to hand it to IBM. There was a time when those guys had a reputation for button-down formality in everything they did. Now I see they've been working drunk for half a decade.

Dig Deeper on Network Infrastructure

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.