WordPress database error: [Got error 134 from table handler]
SELECT * FROM wp_bas_visitors, wp_bas_refer, wp_bas_ua, wp_bas_os WHERE referer = referer_id AND osystem = os_id AND useragent = ua_id AND lasthere > DATE_SUB(NOW(), INTERVAL 20 MINUTE) AND visit_ip = 1679299725 AND ua_string = 'claudebot' ORDER BY lasthere DESC LIMIT 1

The Cargo Cult of Business » A Hacker’s Eye View of Disruptive Computing Technologies

A Hacker’s Eye View of Disruptive Computing Technologies

Published on 8 Mar 2006 at 10:00 am | No Comments | Trackback
Filed under The Cargo Cults of Business, Thanks for Playing, Winners and Losers, Technopolitical, Business and Corporation Related, Networking Technology, Information Technology, Open Source Software, Economics and the Economy.

Remembering SGI 

I was recently in a list group discussion about the recent layoffs at SGI. Whenever I see a former industry titan on the ropes, I can’t help showing my age by rambling on about tech industry history and the endless cycles of disruption that seem to plague it. Since these days I’ve actually got a community blog at my disposal, this time I thought I’d share my mental wanderings with the world. Who knows, if you agree with one of my former managers at Cisco that I "have an amazing ability to predict market trends and directions", maybe you’ll be able to pick up a dollar or two in the stock market. :-)

 [It’s worth noting that since this is an impromptu ramble, I’m doing this all from memory. Bit errors are bound to happen, and of course since this is an opinion piece, YMMV :-]

Speaking with a 20+ year perspective on the technology industry, most of the churn in tech companies seems to be due to the impact of "disruptive" technologies. Now, unless one has actually read Clayton Christiansen’s book, The Innovator’s Dilemma, that statement won’t make much sense since the media quickly seized on "disruptive" as a cool, fly term for "better." But that ain’t any of it, and a pox on the media for scrambling up the discussion of such an important phenomena that we now have to pen paragraph-long discalimers before using the term. ;-) The best way to get a handle on how disruptive technologies work is to read Christiansen’s book:

If you are employed in the tech industry, it would behoove you to grok this concept since it’s extremely useful in determining whether the exit from an employer will be stage right with cashed out stock options, or stage left in a Chapter 13 layoff. But since I want to introduce some additional concepts on top of Christiansen’s dynamic, if you will, I’m going to offer a (somewhat) brief recap of the main concepts. (Christiansen, of course, does a much better job of exposition at the core concept of disruption.)

Tides of Disruption 

Technologies evolve in markets in a trifecta– Good, Fast, Cheap: Pick Any Two. Really mature technologies finally, after various companies have formed to stake their claim to some duo of the trifecta, manage to achieve all three– that’s the very definition of "mature technology." Then all the companies serving a market re-tool for a commodity pricing model, die in the attempt, or take a giant leap of faith and change their whole business mode.

A disruptive technology (say, B) is one which is initially invented to address a duo in market B. But then, as it evolves towards the holy trifecta in market B, it reaches a point (and this may happen early in the evolution) where it turns out to have some moderate utility in Market A (meets one of the three market factors), which is being very profitably served by technology A. Technology A may or may not be in a trifecta state at that point; frequently it is not.

As technology B matures still further, suddenly it turns out that technology B can be used just as easily in place of technology A (now, it meets two of the three market factors). At this point, Technology B has crashed the party and is starting to generate heat against the established players in Market A. But the established players don’t see Technology B as a big threat (even if Technology B has a trifecta in its own space of Market B) because– here it is– it’s not "mature".

But then the apocalypse comes. Yes, you guessed it, suddenly technology B evolves to the point where it hits the trifecta in *Market A*. And, what makes it so "disruptive", is that it’s reached that trifecta either a) before the established players in Market A have or b) it exceeds the trifecta in all areas.

The established players are left reeling; they can’t simply "re-tool"– they’re not in any way set up to provide products for *Market B*, which they’ve never even considered touching before. There is a mad scramble to move "up market", and focus on only the Good and Fast factors, throwing Cheap aside in an effort to keep up a revenue stream.

It never works. The example Christiansen likes is disk drive manufacturers, and with good reason (he says they come and go like fruit flies; if you blink, you’ll miss ‘em). How many 14" hard disk manufacturers are there still alive today? Absolutely zero. How many 8" hard disk manufacturers are there? Again, zero. With 5.25" drives, manufacturers started to  wake up, so there are a scant few left: Seagate, IBM, HP. 3.5" drives were going strong when Christiansen published his book (which made a huge stir back in the late 90’s, and I’m guessing a number of y’all have read it), so most of those companies had their act together and started in on laptop drives and (in the case of IBM) microdrives to freeze out, as best they could, any disruptive start-ups.

Ancient History 

But speaking with the advantage of a longer view, it’s the computer itself that has followed this trajectory better than any other. Now, you could just say that with drives and computers, the stuff simply got "better". But in terms of corporate behavior, that’s not sufficient to explain the mass die-offs. The transistor caused a minor shake-up in the old tube size-of-a-city-block stone age market, but most of the stone age manufacturers were the ones to deploy it in the first place– a case of evolution, not disruption.

The industry enjoyed a golden age for the better part of a decade in the very lucrative business market (Market A), Burroughs, IBM, Honeywall, Control Data, and NCR having reached a trifecta. Various bit players like Pitney-Bowes with their Spice 11 filled out the "Good/Fast" niche for specialized applications.

Then a little company in Maynard, Massachusetts called Digital Equipment Corporation decided to knock up a smaller system using the first rudimentary silicon chips that lacked the massive processing and storage facilities of the big iron, but did exceedingly well at scientific uses (Market B). Thus was the minicomputer (technology B) born, with the duo of "Good/Cheap" for the scientific market.

As the mini evolved, other players like Data General entered the fray, but the mainframe guys continued to scoff at the little Shetland ponies which they could run rings around with the big stuff. But a strange thing happened: The Shetland ponies never grew much in size compared to the mighty Clydesdales, but bit by bit they became more powerful, and they always stayed cheap. Pretty soon, a mini could meet the "Good/Cheap" duo in Market A, and business software began to be developed for them. The mainframers were quick to adopt the silicon chips for their big iron, which gave them the trifecta– for the moment.

By the late 70’s, though, the writing was on the wall. Improved chip manufacturing ability allowed DEC’s VAX-11/780 to pull with the big boys– and at considerably less cost. As the 80’s dawned, the mini reached the Market A trifecta. Mainframe sales went into free fall, and by the late 80’s, IBM– the only one of the big iron companies left– decided to exit stage left into the consulting business. This proved to be a legendary leap of success in changing an entire corporate business model for a huge company. Unfortunately, most companies still resist such radical reinventions. Among those that don’t, getting it right is no mean feat.

Small Is Beautiful 

Of course, we all know what happened to the minis: The lowly microprocessor did unto others as they had done to their forebears. Now, the only mini manufacturer still standing is IBM, and they are effectively selling to niche markets with a sole factor of "Good".

The first wave of microprocessor machines to storm the business world was the graphics workstation we in the industry all know so well. Sun and SGI were the reigning kings, with lesser lights like Apollo and wannabe’s like NeXT and the Johnny-come-lately DEC bringing up the rear.

But the wheel never stops: a simple, low end desktop system was designed for routine business and personal use, rather than high end scientific and graphics, uses.  The workstations used state-of-the-art chips, and advanced coprocessing designs to pump out mini and in some cases mainframe class power. The desktop PC, with its single microprocessor was no match. At first.

Enter the mid 90’s, and the Intel Pentium and PowerPC chips. The same silicon fab tech that made those allowed simple single-chip graphics processors to flourish (Cirrus Logic made their fortunes on the first wave, along with ASICs for Quantum 3.5" disk drives; George knows more about all that than I do, though; I just sysadmined the workstations ;-) .

At the dawn of the millenium, the tipping point was reached, and the PC gained the trifecta in the workstation market. Sun and SGI had pretty much killed off the stragglers before that point (DEC had comitted corporate hari-kari in one of the most spectacular corporate meltdowns ever seen in the industry, sacrificing its installed base as a burnt offering in an attempt to seize the workstation market trifecta. It failed miserably, as did DEC).

From my eagle’s eye view, SGI decided to retreat directly up market in an effort to buy time until something better suggested itself. Trouble was, they already *were* up market. Their prescient CEO had seen the end coming miles off and exited stage right at the high point in a cloud of money. A retreat into software was out of the question; they had maintained a core competency as a platform, and their software resources were limited. Worse, IRIX didn’t have much on Linux, so their was no getting money by trying to port their largely unremarkable (and cantankerous) OS. With nothing to offer in a buyout or merger scenario, they were stuck.

Sun, led by the brilliant Scott McNealy, fared somewhat better, making a dogged push into Managed Services (IBM’s example) and software. Unfortunately, part of their software strategy was a page from Microsoft’s book, and they were relying on OS sales. Open Source operating systems are subsequently eating their lunch, with Open Source software waiting in the wings for a dinner engagement. Hardware sales plummeted, and the company has been slowly bleeding to death since the beginning of the milennium.

IMNHSO, had either SGI or Sun have embraced Open Source at the outset, they would have been able to pull an IBM and shift to a new business model. They could still, conceivably, do this. But see above comments.

Rule 0

So, wither now come the slayers of the mighty PC? Well, you’re holding ‘em. Yep, that’s right, your cell phone and Palm pilot. Think you still need a workstation for the heavy lifting? Think again; all you need is a decent graphics terminal and very fast broadband Internet connection.

You see, there’s another cycle that underlies all the others in the computer industry, one that is driven by basic economies of scale. I call this cycle the "CPU/Bandwidth Engine", and we can postulate it with what I call "Rule 0":

Rule 0: Computing architecture is driven primarily by the cost of of the real-time connections between interactive and processing/storage devices.

That seems obvious enough (well, it does to me anyway), but what this means is that if the cost of connecting an interactive terminal (traditionally, a display, keyboard, and mouse) with processing and storage infrastructure over a distance is high, the industry will favor architectures that place the processing and storage as close to the user as possible. In other words, this is what defines "Fast" and "Cheap".

On the other hand, if the cost of interconnection is low, then the industry will favor architectures that centralize processing and storage infrastructure, because it is much cheaper and easier to maintain that way. This, then, defines "Good and Cheap".

Historically, we’ve seen these two dynamics at work with mainframes and minis, with their remote terminals, and micros, with their desktop ready-access. The Net proved out the interconnection cost, since the whole client-server model is designed to maximize Fast and Cheap over high cost links.

But broadband changes everything. As I said to several Service Provider execs back in the mid 90’s, once you have full-frame- full-motion video and real time input over a remote connection, you’re done. The pendulum will swing inexorably back towards centralized resources.

And we see this today with the ongoing ascendancy of Application Service Providers. As we speak, Microsoft is launching a Net-based suite of its application software. Maintaining local apps is a huge hassle as we all know. Sure, there can be security issues, but there’s no reason we can’t keep our documents local and our apps remote. And our squeamishness about sensitive transmissions on the Net is gone, put to bed by billions of dollars in secure e-commerce.

The Coming Storm

So that’s where we’re headed: Small devices for providing mobility services. And graphics terminals for everything else. Expect to see thin client graphics terminals with local attachment of personal storage devices. Jump drives integrated with the cell phone/organizer and writ large into the hundreds of gigabytes needed to store videos and music– iPods on Arcturan Mega-steroids.

Remote application software for everything from traditional office apps to high power graphic and scientific processing. Expect even video and other large-file activities to go remote; your $24.99 a month ASP cost will have an optional $20.00/month per petabyte storage fee.

Phone numbers will finally vanish (thank God!), replaced with a far more reasonable name/domain combination similar to what we use now. Voice navigation on all mobile devices will make usage of the longer, but more easily managed personal nomenclature acceptable to the older Baby Boomers not familiar with texting and other keypad-text-entry activities.

Wireless will continue to burgeon until finally Net access is so commoditized it rides free (remember, the real money is in application service fees), subsidized by enlightened municipalities or savvy ASP’s. Microsoft will morph, becoming the dominant ASP provider and ceding the desktop thin-client field to open source. Apple will be the reigning king of media and mobility devices, having finally come to their senses and opened up the iPod as a development platform for telephony and mobility apps when cheap foreign mobility platforms powered by open source begin to be a threat). A motley crew of Chinese hardware manufacturers, led by Lenovo, will keep us supplied with thin-client workstations (and those OSS powered iPod wannabes). Dell will grow fat on revenues from high-end central server arrays, with a hand still kept in the thin client workstation and mobility markets. Seagate will have been buried overnight by IBM and Hitachi, who will have successfully brought holographic crystal storage devices to market.

Oh, and after a long and bitter struggle, Hollywood and the other media titans will have finally, kicking and screaming, embraced a brave new world of unheard of wealth and power by removing all but a token layer of DRM from their content (Well, I’m sure hoping to all Heaven that’s how it turns out; Lawrence Lessig does a great job outlining the more likely– and far, far darker– outcome in his book Free Culture: The Nature and Future of Creativity.

So there’s one Hacker’s thoughts on where the computing industry has been and where it’s headed. Invest accordingly. Just remember, you read it here first. :-)

 

-- Paul
Computer Recyclers
Pre-Owned Macintosh Computers, Parts, & Service
Clearance
iMac G4s & G5s
10% off on all PPC imacs in stock during February.

Comments are closed.

Blogroll

Technopolitical

Networking Technology

General Interest

Design, Interface, and Usability

Business and Corporation Related

Blogosphere

Apple Computer Related