LISTEN
To This Issue.
Give those eyes a rest...
Quote of the Week.
Remember when kids would not
give up their TV or telephone?
A Billion Here, A Billion There...
An amazing number of
transistors...
No New Wires.
Powerline networking - a
hands-on review of this latest home & small
office networking choice.
"NOW We're Talking Useful Technology!"
A less pedestrian promise of
nanotechnology and "embedded" man-machine
interfaces...
About "The Harrow Technology Report"
LISTEN To This Issue.
Do you prefer to let your ears do the work of
keeping you in-touch with, and thinking about
where technology is taking us? If so, "The
Harrow Technology Report" is also available in
an audio-on-demand, Web-based, MP3 version.
If you have an MP3 player on your system (and
most do, such as Window's Media Player,
RealPlayer, etc.), clicking on the link below will
either stream the file to you, or, depending on
how your system is configured, it might download
the file before playing it. Alternatively, if you
specifically want to download the file, simply
right-click on the link, and choose "Save Target
As..."
Also, to learn how you can listen at whatever
speed is most comfortable to you, check out the
FAQ at
http://www.theharrowgroup.com/help.htm
.
So, if you wish, just click on the following
link to listen to this issue!
http://www.theharrowgroup.com/articles/20020826/20020826.mp3
.
Back to Table of Contents
Our changing world:
"According to
a study from Knowledge Networks/Statistical
Research, one-third of children ages 8 to 17 say
the Internet is the medium they would choose if
they could only have one, topping television,
telephone and radio.
For boys, television was a strong second choice,
though girls ranked TV third behind the net and
telephone."
Who'd have thought, just a
few years ago...
From
the April 10 edition of
Ken Rutkowski's Daily Tech News Clicks
and eMarketer
http://www.emarketer.com/estatnews/estats/
edemographics/20020410_kids.html
Back to Table of Contents
This picture,

in an Aug. 13 AnandTech article brought to our
attention by reader Sander Olson
(http://www.anandtech.com/showdoc.html?i=1677&p=2),
shows a wafer which is actually about the size of
a pizza. But this high-tech pizza, produced by
Intel using its new 90 nanometer fabrication
process, holds one-third of a trillion
transistors!
The individual chips within the wafer each hold
52 megabits of SRAM cache, and are destined for
new versions of the Pentium.
A billion here, a billion there -- and pretty soon
we're talking REAL chips... (With apologies to
Senator Dirksen, who, it turns out, may not have
actually penned his oft-quoted version of this
saying -
http://www.dirksencenter.org/featuresBillionHere.htm).
Moore's Law continues its march onward!
Back to Table of Contents
Home networks, according to the Aug. 15
Stavance R&D Newsflash, have recently reached the
10.5 million mark worldwide, and are expected to
grow to 16.6 million by 2006.
Many of you have been on the leading edge of
this trend, first pulling cable for 10
megabits/second Ethernet (often called 10bT). Or
you may have dabbled with HomePNA networking (see
below). Or more recently, you may have climbed on
the 10 megabits/second wireless networking
bandwagon under the moniker of WiFi or 802.11b.
But you probably haven't yet experimented with
networking that uses the existing electricity
wires in the wall (the power line) to carry data.
Because power line networking is a relatively new
home and small office option, in this issue we'll
see how it works, and try to put it in perspective
alongside other, more familiar networking
contenders.
Let's begin by reviewing the common home and
small office networking technologies so that we
have a sense of their strengths and weaknesses:
-
10bT (Ethernet, or Category 5) wire
is, for all reasonable purposes, very secure and
delivers a pretty error-free 10 megabits/second of
raw data (or ten-times greater, if your network
cards support 100bT). But you have to pull a wire
to every spot you want a PC. If you use an
"Ethernet switch" at the central point where all
the wires come together, then the full media speed
will be available to each device (a "hub," instead
of a "switch," shares the media's bandwidth among
all PCs). Each PC requires an Ethernet card (if
one isn't built-in), or something like a
USB-Ethernet adapter. Most cable/DSL "modems"
provide an Ethernet jack.
(Note that for all of the network configurations
we'll be discussing in this article, not only do
you need a network adapter on each PC, but you
also need one more "box" if you want to share your
cable/DSL connection -- a "router." It could be
a standalone box, a part of the cable/DSL modem,
or even a PC running the appropriate software.
For this discussion though, we'll assume that the
router is a part of the cable/DSL modem.)
-
HomePNA
(http://www.homepna.org)
piggybacks data onto your existing home telephone
wiring (without interfering with phone calls).
Until recently, it ran at only one to two
megabits/second, but its V2 devices are now rated
at 10 megabits/second. You plug a HomePNA "modem"
between each PC and a phone jack in the wall, and
the PCs are then networked together. If you have
a cable/DSL modem/router, then another HomePNA
modem goes between the cable/DSL modem/router and
a phone jack.
However, according to Intel technical support
(Intel is one of the manufacturers of HomePNA
devices), encryption is not a part of the current
HomePNA products. Because the signals degrade
over distance, this may not be a huge security
risk in a rural area, but in high density
apartment buildings most phone lines show up on a
common "punch block" where unencrypted signals
could be siphoned off.
HomePNA devices are available as PCI cards, USB
adapters, and Ethernet to phoneline "bridges."
(See
http://www.dlink.com/products/DigitalHome/
Hpna/HomePNA-SetupDiagram.htm for one
vendor's example of what a network using a
combination of 10bT wired Ethernet and HomePNA
might look like.)
-
WiFi, or 802.11b
(http://www.weca.net/),
offers 10 megabits/second of data and the ultimate
in flexibility, especially where wandering
notebooks are concerned. You connect a "wireless
access point" to your cable/DSL modem/router via
an Ethernet cable, and then "wirelessly-enable"
each PC or notebook with a built-in PCI card, a
PCMCIA card for notebooks, or a USB adapter. But
depending on the building layout and construction,
its coverage can be spotty, and documented flaws
in the current WEP encryption make it susceptible
to hacking (http://www.pcworld.com/news/article/0,aid,82563,00.asp
and
http://www.weca.net/pdf/20011015_WEP_Security.pdf).
See
http://www.dlink.com/products/DigitalHome/
Wireless/11b/2-4GHzDia.htm for one
vendor's depiction of a wireless network
configuration.
New systems, based on the 802.11a standard
(the reason that the "a" version is the
higher-speed successor to the "b" version is
bureaucracy at work...) raise the raw data rate
from 10 megabits/second to 54 megabits/second.
Unlike the "b" version which shares its radio
spectrum with (and can get interference from)
microwave ovens and "2.4 gigahertz" cordless
phones, the newer 802.11a moves up to the
relatively un-crowded 5 gigahertz spectrum, and it
provides a potentially more robust level of
encryption. Note, though, that these two wireless
standards are NOT compatible with each other. If
you have a notebook with an 802.11b card in it, it
can not talk AT ALL to an 802.11a access point.
If that type of backwards compatibility is
important to you, there are access points that
include both 802.11b and 802.11a functionality.
Which brings us (finally) to the main point of
this discussion -- an introduction to the relative
newcomer to the home and small office networking
arena:
Power Line Networking.
Under the standardizing influence of the
HomePlug Powerline Alliance
(www.homeplug.org),
several companies (http://www.homeplug.org/members/)
now offer products that use the existing
electricity wires within our walls to carry data
at a maximum raw data rate of 14 megabits/second,
without affecting the power.
(One company, PhonexBroadband, explains that the
overhead associated with networking over the very
data-unfriendly power lines reduces that "raw" 14
megabits/second to an effective data throughput of
8 megabits/second (http://www.phonex.com/pdf/Neverwire%2014%20FAQ.pdf),
but more on the speed issue in a moment.) The
recommendation for maximum distance between
powerline modems seems to be about 150 feet,
although HomePlug indicates that they have yet to
run into a house where it didn't work. And a
practical (if not an absolute) limit of 16
powerline modems seems to be what's claimed for
home or small office powerline networks.
Similar to the other technologies, you need a
"powerline modem" between your cable/DSL
modem/router and the wall, and another powerline
modem between each PC and the wall; various
companies make versions that connect to an
existing Ethernet jack, to a PC's USB connector,
or that can be installed into a PCI slot.
I recently had the opportunity to test powerline
modems from three companies, Asoka
(http://www.asokausa.com/prodlist.htm),
PhonexBroadband (http://www.phonex.com/prd_nw14.htm),
and ST&T (http://www.stt.com.tw/PLC/plc-1.htm),
and I must admit that I went into the testing more
than a little skeptical -- my electrical system is
so "noisy" that my few X-10 lights often exhibit a
mind of their own. But -- the powerline modems
did indeed work, and quite well.
Behind The Tests.
Unlike the carefully controlled benchmark-driven
tests conducted by many "testing laboratories"
(which are also valuable), I focus on pragmatic
real-world testing that uses the same type of
equipment you probably have, configured in similar
ways. For example, I don't use PCs stripped of
all of their background tasks to keep normal CPU
activity from affecting test results (although I
do make an effort to see that no unusual activity
takes place during the tests.) And for these file
transfer speed tests, I didn't configure an FTP or
other specialized server. Instead, I used
Window's File Sharing and Explorer's "drag and
drop" to copy a large 338 megabyte (aprox.) test
file between the two systems in the same way you'd
likely transfer a file. Tests were conducted in
both directions and repeated twice, with the
results averaged. So while absolute test numbers
might be better in optimized environments, these
results reflect how I would expect the technology
to actually work for you.
For most tests I used powerline modems from the
same vendor, one attached to my cable/DSL
modem/router and the other to the test PC running
Windows XP; each powerline modem was plugged into
a different electrical circuit (meaning that the
signals had to travel down to the circuit breaker
box and back up.) The second PC in the file
transfer tests was connected by wired Ethernet
directly to the router, and used Network
Neighborhood to gain access to the test PC.
Installation.
Powerline modems that connect directly to an
Ethernet jack, either on a cable/DSL modem/router
or to an existing Ethernet jack on a PC, didn't
require any setup or software install at all --
they were transparent to the machines. Installing
a USB version did require a typical software
installation, which proved flawless on a Windows
XP machine. However, when I originally attempted
to install the ST&T USB software on a "well-used"
Windows Millennium system, the installation
process repeatedly failed and eventually corrupted
the system. This may have been more an indictment
of the state of this old Windows Me machine than
of the powerline modem software, and I didn't
pursue it since this seemed a good time to upgrade
the machine to Windows XP.
J
But all of the powerline modems did work fine
under XP.
The Results.
To begin, I ran a benchmark file transfer between
the two systems using only standard wired 10bT
Ethernet (no powerline modem), resulting in an
end-user throughput of 7.2 megabits/second.
(Don't be surprised that wired Ethernet doesn't
live up to its "raw data rate" promise of 10
megabits/second -- protocol and operating system
overhead always take a toll on the "end-user data
rate," which is why we're going to be comparing
actual end-user file transfer performance.) For
comparison, a previous similar test of 802.11b
wireless Ethernet (high signal strength and only
one wireless PC active) yielded an end-user
throughput of 7.0 megabits/second.
I then installed pairs of powerline modems and
tested the file transfer throughput as described
above, yielding the following end-user data rates:
Asoka PL9610-ETH / PL9710-USB 4.0
megabits/second
PhonexBroadband NeverWire 14 4.6
megabits/second
ST&T M51 / U21
4.9 megabits/second
10bT wired Ethernet only 7.2
megabits/second
802.11b 10 megabits/second wireless 7.0
megabits/second
(You may well experience faster or slower
speeds, depending on your electrical system.)
Analysis.
The most important result of this experiment,
to me, is that these powerline modems provided
exactly what they promised -- simple, no-new-wires
networking.
You might be concerned that the end-user data
throughput of the powerline modems (in my tests)
was slower than wired or wireless Ethernet, but
let's put that in perspective -- if you do intend
to transfer large files or expect to stream lots
of multimedia at speeds faster than your cable/DSL
connection, the increased speed of wired Ethernet
might well justify the hassle of pulling cable.
But for most typical small network activities,
such as sharing a cable or DSL connection,
remote-printing, and typical file transfers, the
slower speeds of the powerline modems may be
unnoticeable compared to the flexibility of being
able to compute anywhere you can plug-in to an
electrical socket.
Security -- Just Say "Yes!"
One caveat, is that out-of-the-box, the
powerline modems all implement a very basic level
of security, which some manuals such as the one
from PhonexBroadband, clearly (and correctly) call "not
adequate protection for your network."
(http://www.phonex.com/pdf/NeverWire14_Users_Guide.pdf)
Happily, each modem came with a mechanism to
implement a more robust 56-bit DES encryption,
although the method to implement it varied.
For the Asoka and ST&T devices, it was trivial to
use a software utility to set your own personal
encryption key for powerline modems that were
directly connected to the PC, either via USB or by
a PC's built-in Ethernet jack. The problem,
however, is if you need to set the encryption key
for a powerline modem that normally connects to
the Ethernet jack of your cable/DSL modem/router,
if your PC does not have an Ethernet jack. You'd
have to install the encryption utility on another
PC that did have an Ethernet jack, temporarily
plug the Ethernet-only powerline modem into it,
and then set the encryption key to match the
others.
This ability to set your own key should allow you
to set the encryption keys on different vendors'
modems to the same value and have them
interoperate with each other. In my case,
however, I was not able to get this to work. It
"should," and it "might" if I'd pursued it
further, but at least initially I'd plan on using
a single vendor for powerline modems until their
interoperability becomes more robust.
Which brings us to PhonexBroadband "NeverWire 14's" method
of setting security -- it's much more elegant. To
set the robust security on the NeverWire, you pick
any one device and hold down its Security button
for 5 seconds. This generates a random key and
puts it into a special "master" mode for five
minutes, during which time you go around to every
other NeverWire and press its Security button for
about 2 seconds -- it then exchanges key
information with the "master" (over the power
line, of course) and gives you a visual indication
of success. Very simple and effective; no
encryption utilities needed, and no special
hookups are needed to set keys on powerline modems
that are not connected to a PC. (If you want to
set the NeverWire 14's key to an explicit value,
PhonexBroadband's Jess Howe advises that the utilities
supplied with other HomePlug powerline modems
should work, and PhonexBroadband may provide their own
utility in the future.)
Power To The People! (OK, it's a bad
pun, but I couldn't help myself...)
Speaking of "power," there's something you
(should) have to UN-learn when plugging a
powerline modem into the electrical socket -- do
NOT plug it into a surge suppressor, UPS, or any
other device that filters the power line! Yes,
that makes sense if you think about it, but if
you've developed the (important) habit of never
plugging sensitive electronics directly into the
wall, that habit might affect your network.
(To find out what
would happen, I initially moved just one of the
two powerline modems onto a surge suppressor, and
the data rate dropped by 66%! When I then plugged
the second powerline modem into a surge
suppressor, the two suppressors filtered out
enough of the powerline modem signal that they
couldn't establish any connection at all.)
The Bottom Line.
The bottom line is that powerline modem
technology works, and better than I had expected.
Yes, it's somewhat slower than some other
techniques, but for many applications the network
speed will not be the bottleneck.
Powerline networking could be even MORE useful
if it were bundled WITHIN computer equipment --
imagine if every PC and related device were
equipped with this by default, so that simply
plugging it into the wall for power also brought
along local networking! (Of course this would be
hard on surge suppressor vendors, as we saw
earlier...)
But Not A Bed Of Roses.
The possible fly I see in the powerline
networking ointment, however, is the number of
years that it has taken powerline networking to
get here (I recall a fair amount of hype about
powerline networking three years ago at COMDEX),
and the tremendous price reductions that have
happened to wireless networking devices during
those years:
-
Today, three PhonexBroadband Neverwire 14
powerline modems (enough to share a cable/DSL
Internet connection with two PCs) costs $387 ($129
each). An equivalent set of three powerline
modems from ST&T would cost $265, or $267 from
Asoka.
-
A similar 10-megabits/second 802.11b
wireless system that I priced today on BestBuy's
Web site, which includes a wireless access point
PLUS router (DI-614+ at $150), a notebook PCMCIA
wireless card (DWL-650+ at $90), and a wireless
PCI card for a desktop (DWL-520+ at $99), costs
only $339.
-
Even a new 52-megabits/second
802.11a wireless setup, consisting of a wireless
access point (DWL-5000AP at $269 after a $60
mail-in rebate (honest, that's what it says at
http://www.bestbuy.com/Detail.asp?m=
488&cat=540&scat=1572&e=11121267), a
compatible notebook PCMCIA wireless card (DWL-A650
at $150), and a compatible wireless PCI card for a
desktop (DWL-A520 at $170) would only cost $589.
I haven't had the opportunity to test the
D-Link 802.11a high-speed wireless cards I priced
above, but I have been very pleased with the
original 802.11b wireless networking which I've
used for years, and so I anticipate that 802.11a
will be even better.
Yet there's still plenty of room for powerline
networking -- as with most technologies, there are
situations where each will shine, and I expect
that the price for powerline modems will drop as
they ramp-up production.
The really good news for us is that we now have
so many, relatively inexpensive home and small
office networking choices! It wasn't too long ago
when powerline networking was a pipe dream and
wireless access points were in the economic
stratosphere. Today, as consumers, we have access
to both.
Happy networking!
Back to Table of Contents
Your
Feedback is Important!
I'd
like to understand your interest in The Harrow
Technology Report, how you make use of it, and
the value you feel it provides to you, your
career, and to your company.
Please send your comments to me at
Jeff@TheHarrowGroup.com .
I
look forward to hearing from you!
And, if you know of other folks who might find
value in "The Harrow Technology Report,"
I'd appreciate your letting them know that they
can subscribe at
http://www.theharrowgroup.com/signup.asp
.
Jeff Harrow
Back to Table of Contents
Finally, following up on our recent exploration
of future histories and where technology may lead
us (http://www.theharrowgroup.com/articles/
20020812/20020812.htm#_Toc16827507),
reader Michael Johnson decided to offer his own
interpretation of how NBIC convergence (Nanotechnology,
Biology and medicine, Information sciences, and
Cognitive sciences) might be put to good use:
"I read the short
paper by Roger Born
(http://writing.borngraphics.com/1future.htm)
with great interest. As I was sharing
the idea with my co-workers it dawned on me that
if you had retinal bots sending signals directly
to your brain, you could program them to "adjust"
the signal coming from, say, your girlfriend, to
look like Demi Moore. An overlay of sorts. She,
in turn could overlay the Tom Cruise program on
you.
Even better, you
could just project your preferred wireless image
to all receivers within range.
This could
eliminate dieting! Go ahead and eat that extra
piece of pie (at least it looks like pie... it
could be cake for all you "know"); no one will be
able to tell.
Today I would like
to be 6 inches taller.
Assuming you had
bots around you auditory system you could even
have the sound adjusted to match your new
physique.
NOW were talking
useful technology!"
I'll bet THAT technology would sell!
About "The Harrow Technology
Report"