For the Week of October 13, 2003
For roughly a half-century the computer industry has been based on the circuit board.
The board defined the connections among components in the final product. At first these were discrete resistors and capacitors, even tubes. But over time these were replaced by transistors and then chips.
PCs were made much as mainframes had been. A board was made, and chips were placed in it, either directly with solder or onto sockets. Memory was here, processing was there. As chips got better, the speed with which the circuit board let signals move between these destinations became a bottleneck. So it was that chips gained "cache memory," an on-chip memory stack that would let processing continue in the foreground while signals moved in-and-out in the background.
Now all that is about to change.
I had an inkling when Sun announced a system whereby it could link chips directly next to one another, without a circuit board . This was confirmed when IBM recently announced a system for piling chips on top of one another .
What this means is that we're leaving the era of the circuit board.
The PC under my desk is based on boards. I can upgrade its capabilities by changing a board, by adding components to a board, or by plugging new boards into the main board or "motherboard." Your cellphone is different. It is a discrete device. When it gives you trouble you toss it and get another cellphone. It's reliable but essentially impossible to fix.
That is about to happen generally. PCs and servers are going to become discrete devices. Their capabilities will be set once, as will their software, and when you want an upgrade you will just replace the whole unit.
Combine this with wireless and you get an entirely new paradigm. A server with a radio on your refrigerator communicates with another server in the wall of your hallway, which may communicate with another one by your desk, perhaps mounted to the back of your monitor. Your keyboard, your speakers, the printer, and all these servers communicate with one another wirelessly, using software defined radios. The server on your fridge may communicate with clients inside using UWB, which can easily go through that metal wall. The server on your monitor may link to your keyboard and printer using Bluetooth. The same device, or any other device within your home, may communicate with the outside world using 802.11. A server on top of a nearby telephone pole, or on your roof, may then communicate with the greater Internet using 802.16.
I have just described a device-based always-on architecture that can not only handle all your current applications, your PC and home entertainment applications, but can also keep track of what's in your kitchen, or in your bath, or can keep track of your health and your grandma. The single-chip clients all have discrete functions - a blood pressure monitor, an RFID chip that knows when your yogurt goes bad, a tag on your keys that lets you triangulate where you hid them, a camera across your door that identifies who is coming in (or trying to), a moisture sensor in your lawn that turns on your sprinkler.
That's the World of Always-On. But what I didn't realize until now is that the entire electronics food chain is going to go through even-bigger changes than your home.
First, "board-bangers" like SCI , and component manufacturing plants like Dell's plant in Austin, are going to go away. Instead, chip designers will work directly with foundries to define, not components, but entire devices. Intel will produce entire products, and perhaps Motorola will produce OEM products if its own brandname becomes worthless. There will be many opportunities in third-party foundries that can take a hardware-software design, replicate it in silicon, package it in plastic, test it, and deliver it to the market with the name of a systems company stamped on the front.
This is a truly profound change, but it's one that will move ahead at a measured pace. Cisco isn't going to suddenly stop making freezer-sized communications boxes. But gradually, more and more of its products will come out as smaller, simpler, use-then-toss components, starting with those it acquired when it bought Linksys and got into the 802.11 business.
This also means that software companies will have to work closely with hardware foundries, and that ISVs will become component designers. It means standards become more vital than ever, but that standards can always be upgraded since there is no sense throwing out the old - just build the new on top of it.
The replacement of PCs based on circuit boards with discrete devices modeled on cellphones is not just a vision for how intelligence will change the way you live. It's also a vision for how the electronics industry itself will change.
These changes will move quickly. My hope is they will move too quickly for Indian and Chinese competitors, because high-end components, like PCs, lose much of their value as they're traveling across the ocean on boats. But regardless, this is where the world is going.
Get with it.
Your list is your most important asset. But what happens when someone forgets who you are and you get on a "spam" blacklist? Your asset becomes worthless.
Need a-clue on how to avoid that? Get your list audited, and managed professionally, by the fine folks at Whitehat , part of the American Computer Group , a long-term leader in database services for direct marketers.
When your list is truly opt-in, not only do you become a white hat yourself, but your e-mails are read, even anticipated, by your audience. That means higher conversions and more money in your pocket.
If you're serious about Internet Commerce, you need Whitehat Interactive . Get it today.
Takes on the News
ICANN declares war on SiteFinder
When something goes wrong with a sports team the fastest way to make things right with the fans is to fire the manager. This is an accepted part of the game. Some teams overdo it, which is why those teams lose. Stability at the top is one sign of a winning team.
In business we judge wins and losses based on sales growth and market share. But the principle should be the same.
Unfortunately it is not. Imagine if, say, New York Mets manager Art Howe controlled all appointments to the Mets' board. Imagine if that board determined Howe's salary as well. Imagine, further, that the Mets were owned, not by one or two people, but by a veritable host of people, none with enough of a stake to threaten that board. Art Howe could easily beat Casey Stengel's old records for losing , and never face the consequences.
That's exactly the situation in corporate America right now. CEOs control the boards, boards scratch their backs with big checks, and the owners have no way to control any of it.
As a result poor performance doesn't get the punishment it deserves. Yes, Chris Galvin of Motorola finally quit, but Michael Eisner's recent record is as bad as the Detroit Tigers - if his were a European soccer club it would have been relegated. Yet he keeps getting richer.
Richard Grasso didn't invent this system. He merely took full advantage of it, becoming incredibly wealthy for a job that was supposed to be that of a figurehead.
All this changes only when the fans, in this case stockholders, get into the game and demand value for money. Big salaries are fine, but only if we can fire you in a heartbeat.
ICANN declares war on SiteFinder
ICANN's decision to call Verisign on its SiteFinder "service" is long overdue. The question is whether it will do any good.
Verisign backed down temporarily but cast itself in the role of victim, continuing to call its attack on the Internet infratructure a "service" (a line that was meekly accepted by most reporters). The company also said it was reviewing its legal options, and that may be the most ominous point of all.
The sad fact is that the U.S. Department of Commerce, and therefore the U.S. President, control the Internet. ICANN itself is merely a fig leaf over this fact. Verisign has figured this out, and its decision to launch SiteFinder must have been taken with this in mind.
The time has come for the international community to start taking Internet governance seriously. Once Verisign rips away the fig leaf , it will be up to ICANN's international members to take back the power, first by acknowledging their powerlessness, then by bringing the Internet under true international authority.
Dangers In Consolidation
U.S. technology businesses have two primary advantages against foreign competitors.
One is speed to market, illustrated by Dell Computer, which took over the PC business in the 1990s by being closer to its buyers than either Japanese or Taiwanese manufacturers.
The other advantage held by U.S. companies is innovation. By innovating rapidly, U.S. companies forced foreign competitors to open offices here, and follow U.S. standards, starting with the IBM PC.
This had its ultimate expression in the Internet boom. Sun Microsystems, which had already built a big business in engineering workstations before the Internet got hot, dominated the era with its servers and Java language - it became "the dot in dot com."
Hewlett-Packard's recent announcement of a $25,000 bounty on Sun customers, for its Linux servers,
represents a rejection of this direction. H-P CEO Carly Fiorina sees the hardware and software businesses as consolidating. There will be fewer, bigger players. She figures that, by buying Compaq and squeezing out smaller companies (like Sun ) she could guarantee H-P's place in that new universe.
She can't. The result of the recent consolidation is to slow innovation and, thus, the advantages in being quick to the market. Dell's recent move into consumer electronics - buying massive quantities of products from overseas - is another acknowledgement of this trend.
This trend benefits only China. Most Intel fabs are already located closer to Beijing than to Palo Alto. With speed-to-market not a factor, and innovation slowed to a crawl, Chinese manufacturers now have a chance to not only kill Sun, but H-P and Dell as well.
Clued-in is the move to reduce fines for file trading . Even the industry's supporters should agree that fines of $150,000 just create martyrs. More clued-in would be a valid, market alternative to file trading that would support both the consumers' and artists' interests.
Clueless (unfortunately) is the consumer suit against Microsoft over its security lapses . The solution is the West Coast Law of competition, not the East Coast Law found in a courthouse.
A-Clue.Com is a free email publication, registered with the U.S. Copyright
Office as number TXu 888-819. We're on the Web at http://www.a-clue.com.