Nov. 10, 1997
"Snap! Crackle! -- Click?"
by Jeffrey R. Harrow
Senior Consulting Engineer
Corporate Strategy & Technology,
Digital Equipment Corporation
jeff.harrow@digital.com
Insight, analysis and
commentary on contemporary computing
and the technologies that drive it (not necessarily the views
of Digital Equipment Corporation).
Copyright ©1997, Digital Equipment Corporation
I'd like to understand your interest in the RCFoC, how you make use of it, and the value you feel it provides to you, your career, and to your company.
Please send me your comments to Internet mail address:
jeff.harrow@digital.com
I look forward to hearing from you!
Jeff Harrow
If you'd like to give your eyes a rest while keeping up on the innovations and trends of contemporary computing, then "RCFoC Radio" is for you. It works if you're behind a firewall, and it works if you're using a 14.4 modem. So jump in and test the "new media" waters. Point your Web browser to http://www.digital.com/rcfoc/audio/971110.htm to experience the sounds of the Rapidly Changing Face of Computing.
Today, the silicon crystals that are sliced into thin wafers to form the basis for our chips are "grown" under very carefully controlled conditions. But the resulting wafer must then go through many complex and exacting processes in fabulously expensive "fabrication plants," or "fabs," to precisely lay down the multiple layers of circuitry needed for the final result. Wouldn't it be easier if we could just "grow" the finished chips -- circuitry and all?
"Absurd!" "Impossible!" "Ridiculous!"
Yet, that's the end-goal as Yale University researchers are working to convince organic molecules to form themselves up into the tiny wires and semiconductors that could pack billions of transistors on a single chip!!
Of course this is just some of the very preliminary work that could, eventually, lead to such a result. Specifically, according to the Oct. 10 Science article "Conductance of a Molecular Junction," the Yale team measured a current through an organic benzene molecule they convinced to line itself up between two gold electrodes. This formed a "statically stable gold-sulfur-aryl-sulfur-gold system, allowing for direct observation of charge transport through the molecule," -- got that?
Well, happily most of us don't have to understand those details any more than we really know the details of how today's chips are manufactured. But if this tiny seed of knowledge were to grow into the desired end result, molecular-scale electronics could change the face of our businesses, and indeed our society, the same way the transistor did years ago.
Why? Because, according to the Nov. 3 Business Week, should this come to pass, one laboratory beaker could "hold more transistors than the chip industry has produced over the past 30-odd years." And we'd have lots of beakers...
"Here there be dragons." Potential revolutions that could, indeed, change ALL the rules.
Technology opens many doors, some good and some bad. One element that straddles both sides of that fence is "privacy." Quite rightly, a lot has been written and discussed about how instant access to information can be abused (such as automobile license plate data being instantly accessible to "stalkers" from an Internet-connected mobile notebook). But there are times when, with appropriate controls, technology that can track people's movements can benefit society as well.
According to the Nov. 3 Computergram, Airtouch Cellular recently responded to a court subpoena by using their cellular network to track the movements of a suspected criminal. How?
The cellular network is composed of many towers spread in a mesh across a city; each tower communicates with cellular phones in its vicinity. As you move around, the cellular network notes as you're departing one cell's area and figures out which adjacent cell you're entering; it then transparently switches you over to the new cell -- hence the cellular network as a whole can, to a degree, track your movements.
Now it probably doesn't surprise you that this can occur while you're talking on the cellular phone, but did you know that in most cases it happens any time the phone is turned on, even if you're not making a call? It's because, in order to be able to deliver calls to you, the cellular network and your phone periodically have a brief quiet chat so that the network knows you're ON and where you are...
And these techniques aren't limited to the U.S. cellular systems -- the UK recently used a similar method to track people involved in a soccer "goal-fixing" scheme, and the evidence was used in court.
But what does this have to do with computing? Such techniques aren't limited to pocket phones. With the expanding use of data-over-cellular, and with the coming of Internet In The Sky within five years, the number of communicating computing appliances that you may casually drop into your briefcase or pocket just might, unexpectedly, reach out and touch more than you bargained for...
As I said, every technology carries the seeds for both good and bad; as technology continues to march forward each innovation will offer its own unique opportunities for value, and for abuse. That's not a reason to shun the technology. But each of us KEEPING ABREAST of the rapidly changing face of computing is one way to KEEP the rapidly changing face of computing -- and how it is used -- under control.
Speaking of changes, we recently touched on one of the revolutionary sciences that holds the potential to change ALL the rules, from computing to medicine to construction -- nanotechnology, the science of the tiny (http://www.digital.com/rcfoc/971020.htm#Tiny).
It's worth knowing a bit more about this. And with thanks to RCFoC reader Thom Rogers for bringing this to our attention, Paul Saffo offers a thoughtful, enlightening, and understandable view of some of the consequences we may be seeing from our growing ability to make smaller machines to do our tiniest bidding.
Paul makes the case that improving nanotechnology will be impacting everything from sensors that already help our computers stay aware of our environment (such as tiny radar chips that could automate an oil dipstick), to tiny machines that could allow our computers to directly manipulate our environment, to the development of smart materials and intelligent artifacts -- "smartifacts." Notice the trend? From "sensor" to "effector." And that trend has enormous implications,
"Now we are handing sensory organs and manipulators to the machines and inviting them to enter analog reality. The scale of possible surprise that this may generate over the next several decades as sensors, lasers, and microprocessors co-evolve is breathtakingly uncertain."
You can, perhaps, recover your breath, with Paul's help, at http://www.iftf.org/sensors/sensors.html . This one -will- get you thinking, about the rapidly changing face of computing.
No, this isn't a patriotic slogan by some Technology Party, but the latest move by MasterCard to get Smart Cards into the U.S. market. We're all pretty familiar with the magnetic strip technology that has been flaking off of our credit cards for years -- the strip carries a digital version of our name, account number, expiration date, and a bit more. But in general, these stripes are a "write once" medium and they aren't altered when they're "swiped." Smart Cards (http://www.mastercard.com/smartcard/), however, don't store information this way -- they have a tiny microprocessor and memory embedded within the card to remember static data, and to let new encrypted data, such as cash balances, be written and re-written. Depending on the card's design, they are powered and communicate with the outside world either through that golden design on the surface of the card (they're physical contact pads that are touched by the card reader), or by a low powered "contact-less" radio signal.
Smart Cards seem like a good idea, and they have lots of potential for helping Ecommerce evolve, such as supporting high-volume, low-value "cash" transactions that the traditional credit card transaction network can't support. And, they may make our lives less plastic-ridden -- these intelligent cards can potentially hold multiple "applications." A Smart Card could, for example, be a MasterCard one moment, an ATM card the next, and then a "cash card." And these cards may also contain completely new classes of functionality, such as "loyalty applications," where a store might reward you for being a good customer.
If you live in Europe, your wallet probably already computes in this manner, but Smart Cards haven't yet been widely deployed in the U.S. However, that's now beginning to change. Tests were conducted at the Atlanta Olympics, and another is currently underway in New York's Upper West Side -- check out http://www.mastercard.com/smartcard/map.html for an insight into what these early cards provide.
MasterCard is preparing for widespread Smart Card support; and they've begun beta testing a new set of Smart Card services. These include the card's multi-application operating system (MultiOS) which contains built-in "firewalls" to prevent on-card applications from communicating with each other, the MultiOS Extensible Language for writing these on-chip applications, and, shortly, support for the JavaCard API (http://www.javasoft.com/products/javacard/index.html).
If you're intrigued, there's lots of information available; a couple of places to begin might be the Smart Card Industry Association at http://www.scia.org/ , and CardTechnology magazine at http://cardtech.faulknergray.com/ .
What can we expect to see? During the transition period, as vendors and financial institutions gear up with the new Smart Card readers, the Oct. 31 ZDNet (http://www.zdnet.com/zdnn/content/zdnn/1031/200156.html) suggests that Smart Cards will likely support both the old stripes and the new chips. And although it may take some time, I expect that Smart Cards are one evolution that will surely take place. A meeting of the old and the new, with the potential to dramatically change how we go about our daily business, driven by the rapidly changing face of computing.
ISDN, once it's installed and set up (often one of those "non-trivial tasks"), is a very nice step up from even a 56K modem, in part because each ISDN line actually contains three logical channels. One is a slow (16 Kbits/second) channel called the "D channel." It's primarily used to let the ISDN card or box discuss control of the other ISDN channels with the phone company. (However, in some areas, this is also used as an "always on" low speed Internet connection called "Always On/Dynamic ISDN" (AO/DI) - http://www.pacbell.com/products/business/fastrak/networking/isdn/info/mag/97-summer-whatsnew.html .) The other two channels are identical high-speed "B channels," each of which typically runs at 64 Kbits/second.
Now a nice clean 64 Kbits/second channel, which can also accommodate compression, is pretty good by itself. But it's also possible to "bond" both B channels into a single 128 Kbits/second (before compression) channel that really does leave the fastest analog modem in the digital dust.
The thing is, you can't always establish that bonded connection because each ISDN B channel might end up connecting to your Internet Service Provider through a different router; not all routers can resolve such "split" connections between themselves, leaving you out in the 64K cold.
(Note that "bonding" is also headed for the world of analog modems. A recent announcement from Diamond Multimedia, known for its Supra line of modems, describes their multilink protocol "Shotgun" (http://www.pcworld.com/cgi-bin/database/body.pl?ID=971103170029) that can provide as much as 105 Kbits/second downstream and 67 Kbits/second upstream (uncompressed) data using analog modems. It does this by connecting two 56K modems, over two regular phone lines, to your ISP, then combining or "bonding" them together into a single higher-speed pipe. Call Waiting on the second line allows the data connection to revert to a single channel when there's an incoming call. Of course your ISP must choose to support a single user tying up two lines...)
Bonding, whether with your ISDN line, your kid, or your significant other, is a Good Thing. If you're interested in learning more about the first instance, check out a recent article by ZDNet at http://www.zdnet.com/cshopper/content/9711/cshp0114.html . If you'd like to pursue the other two forms of bonding, I leave that up to you...
Sesame Street's Cookie Monster might be aghast, but some people are tossing their cookies (at least figuratively). It's not due to worries over cholesterol or sugar, but of security.
"Cookies," in the context of Web browsers, are small files that a Web site can ask your browser to create on your disk to hold certain persistent information about your visit (meaning it remains available the next time you connect to that Web site). This can include information such as your account identifier (for those sites you 'log into'), your areas of interest (based on what you did last time), which ads the site showed you (so it can choose the next ones to display), personalized setting you may want to show up each time you connect to that Web site (such as the types of news stories that interest you), or what products you were thinking about ordering on your last visit.
As you can see, cookie files don't come from information you otherwise have on your disk, but only hold information provided by the Web site you're connecting to (which can include information that you provided to that site through your online transactions.) Each Web site can store up to 20 cookies, each containing a maximum of 4,096 characters. And cookies "expire" and are deleted after a defined time.
But still, you might not want information you presented to one Web site to be available to another site. Nor might you want to even admit to one Web site that you had visited another site at all! Unlike your mother's chiding to "share your cookies with your friends" however, a correctly working browser should only disclose a cookie's information to the same site that originally created it.
Of course, there's more to baking and eating Web cookies than I've summarized here, but Jeff Carlson offers an easy to read tutorial on cookies and their crumbs called "Toss Your Cookies" in the Oct. 30 NetBITS (http://www.netbits.net/nb-issues/NetBITS-006.html#lnk2).
Me? I like my cookies warm right from the oven, but that doesn't keep me from letting my Web browser store the digital version that makes my Web experience a bit more personal (although most browsers do allow you to turn cookies off).
It's all a part of the digital give and take in the rapidly changing face of computing...
Let's see. The last time I looked at the price of storage was in September when Western Digital announced their Caviar 6.4 Gbyte drive for $429. That brought the price of storage down to 6.7 cents/Mbyte. But a recent announcement makes 6.7 cents/Mbyte seem expensive:
Quantum has just announced their 12 Gbyte Bigfoot TX, available later this month for -- are you ready -- $399!! That's 3.3 cents/Mbyte!
That's halving the price of storage in less than two months! (http://www.quantum.com/corporate/pc/txpresskit/)
Forbes magazine recently noted that, "the average personal computer sold this year will have ten times the storage of just four years ago" (http://www.quantum.com/corporate/comm/pr/bigfootlaunch.html). Given that the cost of storage plummeted from 6.7 to 3.3 cents/Mbyte in just two months, what might storage cost four years from now? How soon will we have to switch counting the cost of disk space from cents/megabyte to cents/gigabyte?
By the way, just in case you think that these inexpensive multi-gigabyte drives mean that no one will ever again have to think twice about having enough storage, consider what one of the leading-edge consumers of storage is up to.
Switzerland's CERN particle research lab is working on the Large Hadron Collider (http://wwwlhc01.cern.ch/), a particle accelerator due to be operational by 2005 which will generate test data at the rate of 100 terabytes (that's one hundred thousand gigabytes) per second, of which .1 gigabytes will have to be stored. For every second of each test. That's expected to be as much as 5 petabytes (five thousand terabytes, or five million gigabytes) of data per year, according to the Fall Forefront magazine...
We are always pushing the limits. We've never stopped, and I suspect we never will. Storage -- one good example of the (very) rapidly changing face of computing.
Well, now that we're into petabytes for storage, RCFoC readers Dave Nentarz and Jay Cruz suggests that we may soon have to consider petabits for our networks! According to an article by BBN's Craig Partridge in the Oct. 21 Data Communications (http://www.data.com/25years/petabit.html), ISP traffic is currently increasing between 15 and 30 percent per month -- "...stretch that over 12 months and annual growth rates are 500 percent to 1,800 percent."
If Internet-over-cable-TV continues its spread, and if more pervasive and less expensive ADSL and related technologies bring higher speed data to anyone who wants it (http://www.digital.com/rcfoc/971013.htm#The_Connectivity_Choices_They), petabit-per-second (that's billions of megabits per second) Internet backbones could be needed within five years. In fact, the article points out that researchers are already tossing around the rather arcane terms of exabit (a thousand petabits), zettabit (a thousand exabits), and yottabit (a thousand zettabits).
Sounds improbable? But remember that the term gigabit was just as arcane to most of us only a few years ago!
Those capacities are mind boggling -- and Data Communications points out that nobody knows how we might achieve them (at least not this week). But they also suggest that fiber holds some very powerful potentials. One is in expanding our current use of 16 channel "Wavelength Division Multiplexing" (WDM), where one fiber can contain potentially thousands of parallel high-speed data streams. Each data stream uses a different color, or wavelength of light to keep it separated from its neighbor, up to a theoretical limit of about 75 terabits/second of data per fiber.
The other promising technique builds on the way that analog modems routinely pack more than one bit of information per cycle -- by varying the strength of the signal at different times during that single cycle. Researchers are already exploring sending 3 bits of data per pulse of light, tripling fiber's theoretical limit, and we should have learned from analog modems' progression from 2400 to 4800 to 9600 to 14,400 to 28,800 to 33,600 to 56,000 bits/second that we're never satisfied to stop pushing...
Switching such vast amounts of data, when it gets to a fiber crossroads, is another issue, but Stanford University is working on "Tinytera" (http://tiny-tera.stanford.edu/tiny-tera/index.html), a soda can sized switch that can handle a terabit/second, many times the capacity of today's commercial switches. And routers capable of handling terabits/second are on the way...
Of course that "mere" terabit switching and routing speed only gets us started along the way to petabit speeds -- the article suggests that a combination of Moore's Law's faster processors, better engineering techniques, improvements in routing table management, and parallelism may, eventually, get us the rest of the way...
Will we need such networking capacity? Only a couple of years ago a vast majority of the traffic on a LAN was destined for another computer on that same LAN -- today, UCLA research finds that 90% of LAN traffic is leaving the LAN -- to be carried by the Wide Area Network whose capacity we've been discussing. Research at MIT also offers a tantalizing glimpse of wireless 200 Mbits/user connectivity, which could spawn a vast number of data-generating, communicating computing appliances.
Will we one day be tossing around "yottabits" as casually as we toss megabits today? According to Craig, "Even if growth continues to boom, we won't get to yottabits for about 15 years." So rest easy. For awhile. A little while...
Could the days of the "flat" TV be near an end? No, R2D2's holographic projector isn't waiting at your local consumer electronics outlet (yet), but at least two companies are getting ready to use the fruits of Moore's Law's ever-faster processing power to turn 2-D into 3-D.
Chequemate Third Dimension (http://www.c-3d.com/www-html/c-3d/about.html) is offering a $599 box that connects as the last device in the video chain between your VCR, cable box, etc. and your TV. It digitizes the video, and in real-time alters it to simulate 3-D information (called "3-D Conversion"), displaying two slightly offset images on the TV every 30th of a second. You wear "LCD shutter" glasses that sync up with these images so that each eye sees only the "correct" image. The result is a 3-D "effect" (not actual 3-D) that can be applied to any standard (NTSC and PAL) video signal, including video games and, I would expect, to computer-generated video. The 3-D effect gets even better when viewing a video source explicitly encoded with actual multi-camera information; they call this mode "3-D Display" and a few videotapes, such as the one they include, are currently encoded in this manner.
The Sept. 25 Nihon Keizai Shimbun also describes what appears to be a similar technology (but not yet a product) announced by Sanyo Electric; it determines how to simulate the 3-D distance between objects by keying off of the shading and composition in the image.
I haven't seen Chequemate's "C-3D" yet but, having seen multi-camera 3-D images through LCD shutter glasses in the past, these "3-D Display" images should be quite compelling (the same technique can be used for business data visualization and scientific computing tasks). What I do wonder about is the simulation techniques, which would have to be pretty good to not seem artificial and disturbing. We'll see (sorry.)
But in any event, even if these initial 3-D simulation products don't deliver the quality we'd like, Moore's Law dictates that every 18 months we'll have twice the processing power to throw at the problem, and engineers will be developing ever-better algorithms to consume it. Could the worlds behind our video screens cease to be flat? Tune in next {week | month | year} to the rapidly changing face of computing...
The choices for accessing the Internet on your own very mobile terms, rather than tethered by wires or even by conventional "computing devices," are increasing.
For example, consider a new $12/month service from E-Now that allows you to "dial into" your standard POP3 Internet mailbox -- from your phone. You call the E-Now 800 number and give it your ID code. It logs itself in to your mailbox and then uses text-to-speech software to read you a summary of your messages, and then the body of messages you select. You can also access a live operator who can dash off that important reply, fax you the details, or perform other Email related tasks (for which there are additional charges).
There are, of course limitations. Your Email server can't be on the private side of a firewall (as are most corporate servers), and the server must use the POP3 protocol (which currently leaves out AOL and CompuServe mailboxes). Check out 888-hear-e-now for a demo, or find more details at http://www.enow.com/html/main.html .
Of course, if you'd prefer to use your notebook for Email, Internet browsing, or anything else but don't want to tether it to a phone cord, you could hook it up to your cell phone.
Typically, though, dealing with the notebook, cables, adapters, and the cell phone makes for a bit of a juggling act. Now, however, GlobeWave has introduced a Type III (thick) PCMCIA card that isn't just a modem, but is the complete analog modem plus cellular phone! A small antenna sticks up from the outside edge, where you can also plug in a tiny earphone/microphone for voice calls. It's integrated into Windows 95's dial-up networking and, assuming that you're in a reasonably good cellular area, it should "just work." The "Com.plete PC Card's" 14.4 modem also works as a standard wired modem (http://www.globewave.com).
Of course if you want to be able to "drive the Internet" with this mobile access, you might need a car to support your mobile surfing activities.
According to the Oct. 24 Internet Daily, GM has recognized that the "mobile Internet" is, er, a two-way street; they're letting folks in California determine which cars are on dealer lots, schedule test drives, and determine, and hold, the best price right online (http://www.gmbuypower.com/).
I loved this quote from their Web site - "The dealer's best price means there's no need to haggle." Talk about the Internet, and the rapidly changing face of computing, improving our lives!
Finally, few of us are surprised anymore to find that software, in one form or another, is invading not only our business but also our personal lives. But perhaps the breakfast table is going just a bit too far (so to speak)?
Yet if you start your day with cereal from Kellogg's, the Oct. 31 Internet Daily tells us that soon 80 million cereal boxes will sport ads and discount coupons for software, joy sticks, and other game controllers. I can just visualize Tony The Tiger playing the latest PC game and roaring "It's GRRREAT!"
Could a higher-tech, PDA-touting Rice Krispies' "Snap! Crackle! Click!" gang be far behind?
"The Rapidly Changing Face of Computing" is a weekly technology journal providing insight, analysis and commentary on contemporary computing and the technologies that drive it.
The RCFoC is written by Jeffrey R.
Harrow (jeff.harrow@digital.com), a
Senior Consulting Engineer for the Corporate Strategy &
Technology Group of Digital Equipment Corporation.
The RCFoC is published as a service of, but not
necessarily reflecting the opinions of, Digital Equipment
Corporation. Copyright © 1997, Digital Equipment Corporation.
All rights reserved.
Where To Find the
|
AVAILABLE TO THE PUBLIC AT NO CHARGE
subscribe rapidly-changing-face-of-computing YourEmailAddress
subscribe rapidly-changing-face-of-computing-notify YourEmailAddress
|
Creation Date: 07 December 1997
Legal