Home Networking Prepares for Prime Time

With Intel behind it, what could go wrong? Well...
Publish date:

A couple of months ago I

dissected the fast-Web-access market here, looking at the technology, the players and the likely outcomes. Today I want to take a similar whack at the emerging "home networking" market.

The parallels between the two are striking: Several incompatible technologies, with deceptive claims made for each; a number of untested vendors, some with little experience in the field; pricing that's hard to make sense of, high or low; and few guideposts for investors trying to figure how to make a buck.

Both markets have been in pause mode, waiting for galvanizing moments, those big announcements of new products, new pricing or other dramatic incidents that become defining moments, spurring on sales by an order of magnitude. But neither's going to get one; it isn't that simple.

Yet this is such a rich market --

Forrester Research

forecasts a $1 billion annual home-LAN market by 2002 -- that it's tough to sit it out till the smoke clears.

Good news: Tuesday


(INTC) - Get Report

will announce that it's jumping into the home-networking business with a product line and aggressive pricing that's likely to help validate this fragmented and confusing market, and push a lot of us to start networking our home PCs, printers and modems. Unlike a lot of claimed market-validation moves, however -- which rarely do much more than attract some quick headlines, leaving worried investors still whistling past the graveyard -- this one probably will have an effect.

With Intel behind it, the thinking goes, what could go wrong? It must work. Yes? Please?

The Technology

If you want to network your home PCs, and they aren't sitting right next to each other (in which case you could use a simple cable and Windows' built-in direct-connect services), you have four choices:

  • Wire the house with Category 5 UTP (unshielded twisted-pair) telephone-style cable, for a true, full-speed standard Ethernet wired network. This is a mess, and will probably be expensive, as your cable installer has to open some walls, spend a long time crawling around behind things, and tell you at least five times that he thinks this may not work. (I'm assuming you and the family are not charmed by the idea of loose wire lying about everywhere on the floor, running along baseboards and over doors.) Add an inexpensive D-Link or NetGear network hub for as little as $75 or so, drop a $30 network card into each connected PC (make that $150 or so for the special PC cards required for notebooks), and get your network contractor -- you don't want to do this yourself -- bring up Windows 98 on each machine, in a file-sharing mode. (You could also take the techno-heroic route, buying a network server for $2,500-$7,500, running Windows NT on it and maybe on some of the PCs as well, but that's nuts, wild overkill for a home network.) Advantage: You can use either 10 megabits per second or 100Mbps Ethernet, for more speed than you'll ever need. (Tip: Don't opt for the older, lower-speed 10Mbps standard, which is just about obsolete; the cost delta for 100Mbps is so low that you'd be foolish not to go for it.) You also get relatively high security and reliability. Disadvantage: Cost, complexity and, probably, maintenance. Expect, depending on your setup and how much work you try to do yourself (read: your threshold of pain) to spend about $200-$400 per connected device, including the networking parts, installation of the cable, and getting everything working right.
  • Connect your simple home network over a wireless network. This is the simple and relatively elegant answer, and I like this path a lot. You can get a wireless networking card for about $150 for desktop PCs, wireless net LAN cards for notebooks for about $200, and a shared wireless hub/modem for maybe $300. (If you want to share a high-speed cable or DSL Web connection, you'll spend another $400 for a wireless Ethernet bridge.) No wires, no wiring contractor, no holes in the wall. You'll still want some technical help getting this working together, but that shouldn't run more than $50 per connected device. Proxim's (PROX) Symphony is the best of the several wireless home-LAN systems available. Look for Lucent (LU) to bring out a scaled-down home version of its excellent but much more expensive office-level WaveLAN system later this year. And Diamond Multimedia (DIMD) also makes a wireless kit, but its performance is mediocre. Proxim's one-year jump on the competition in wireless home networking, and its 15 years in the wireless business, are proving a powerful advantage. Advantage: Quick, clean, relatively inexpensive, reliable performance. Disadvantage: You'll only get about 1.5-Mbps performance, or less -- a seventh of "old" 10-Mbps Ethernet's speed, and less than 1/60th the speed of 100-Mbps Fast Ethernet. But let's put that in context: You'll still be shipping information around awfully fast at 1.5 Mbps, and I don't think you'll miss a thing, unless you're sharing huge databases, or trying to run video over the network ---in which case you need a big, honking network far faster and more expensive than this market envisions. (Quick rule of thumb: 1.5 Mbps is about twice the speed at which your PC transfers data to a floppy disk.) Moreover, you'll love the convenience of walking outside with a notebook, or watching the 11 p.m. news in bed with your notebook on you knees as you check email ... always connected but untethered. There are some security risks here -- any signal I put in the air, you can probably intercept and read, with enough time and effort -- but the best of these systems use spread-spectrum frequency-hopping (don't ask) in the 2.4-GHz range, which is pretty hard to intercept and reassemble.
  • Run the network over your existing two-pair phone wiring. This is another compromised-speed answer, but it's pretty easy to set up, and it can be very economical. Diamond Multimedia's HomeFree Phoneline Desktop kit, with cards and software for networking two PCs, sells for about $99; additional PCs will add about $50 each. Tut Systems' (TUTS) HomeRun kit is also a phoneline network. Conventional internal HomeRun cards run about $150; external devices are about $175. And Tut has partnered with Broadcom (BRCM) on research to produce new phoneline-networking chipsets with speeds up to 10 Mbps soon, they say, and maybe as much as 60 Mbps someday. And Intel's new AnyPoint system, due Tuesday, uses home phone wiring, too; it's expected to run about $100 per connected PC. (Would you pay a premium for the comfort factor of that Intel brand? Of course you would.) Phoneline networks typically deliver around 1 Mbps, which isn't bad. And they're easy to set up. You won't want to network more than four or five PCs this way, but within that limitation -- how many PCs do you have in your house, anyway?!? -- they're good answers. Advantages: economy, ease of setup. Disadvantages: Limited performance, limited expandability, possible interference.
  • Finally, you can use your existing AC power wiring, already in your walls, for a powerline network. This is a little dodgy -- people have been selling powerline networking for more than a decade, with some dramatic failures -- but today's powerline systems, mainly the Intelogis PassPort system, are safer and more reliable. These powerline nets send a low-power, relatively low-speed signal across your home's wiring grid. Adapters plugged into the parallel ports of each connected PC intercept the signal and turn it into something your computer can understand. An advantage here is that there's no need to pop the hood on your computer to set up the network: Just plug it in and go. On the other hand, powerline networking is sloooow. You're limited to maybe 300 Kbps, which is too slow for most systems, especially as you come to rely more and more on your home LAN. At about $100 per connected PC, powerline networking is relatively inexpensive, but still not a very good value. Advantages: Easy, no-tools installation; relatively inexpensive. Disadvantages: Slow, subject to powerline noise and interruptions.

So who wins?

Ultimately, this is all going wireless, but not via Proxim's present, proprietary technology. Wireless systems compliant with the latest IEEE 802.11 standard will take over. Present systems cannot be converted to 802.11 operation, but at least retooling for the new standard is relatively easy for vendors.

Near-term, the phoneline vendors look like the big winners, with Intel, Diamond Multimedia and Tut Systems leading the pack. Prices will move down fairly quickly, and I think Intel may be able to bump speed up faster than its competition expects.

Conventional wired networks make a lot of sense in new-home construction (where a fiber-optic line is also a good addition, for future distribution of high-speed, high-bandwidth video signals), but the mess of pulling network wiring through the walls of existing home is rarely worth its cost.

And I have little hope for powerline networks: too many problems, too little performance.

There are two big and unstoppable drivers of the home networking market: The rapid growth of multiple-PC homes (the percentage of two-or-more-PC homes is expected to double from last year's 12% to more than 24% in 2001, according to researchers at the

Yankee Group

); and the explosion of homes with fast access to the Web (Yankee says that market will grow 15-fold by 2002) who want to share that access among several PCs.

Intel's timing is just about perfect, jumping into the phone line networking market as it takes off. And Proxim and Lucent are likely to hold onto their niche markets for wireless nets in the home. Everyone else, outta the pool.

Jim Seymour is president of Seymour Group, an information-strategies consulting firm working with corporate clients in the U.S., Europe and Asia, and a longtime columnist for PC Magazine. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. At time of publication, neither Seymour nor Seymour Group held positions in the companies discussed in this column, although positions can change at any time. Seymour does not write about companies that are consulting clients of Seymour Group, or have been in recent years. While Seymour cannot provide investment advice or recommendations, he invites your feedback at