SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : AUTOHOME, Inc -- Ignore unavailable to you. Want to Upgrade?


To: Jay Lowe who wrote (4308)1/15/1999 3:42:00 PM
From: ftth  Read Replies (4) | Respond to of 29970
 
15:13 ET Market Rumors: Wall Street analysts and traders actively discussing possibility of America Online (AOL 147 1/4 +2 3/4)/Hughes Electronics (GMH 47 1/4 +3 3/8) joint venture. According to sources, the two companies would potentially develop a competing service to AtHome (ATHM 103 7/16 +13/16). Rumors bolstered by announcement this morning by AOL of deal with Bell Atlantic to provide high- speed Digital Subscriber Line (DSL) access to AOL, for an additional $20 per month.

and

thestreet.com

Winning in the Fast-Access Wars, Part 1
By Jim Seymour
Special to TheStreet.com
1/15/99 12:42 PM ET

One of the big tech stories in 1999 will be advances in the fast-Internet-access-to-homes business. Unfortunately, the story is complex and hopelessly confused, full of overclaims, techno-speak and hidden agendas. I could write a book untangling the technology, claims and likely market response -- a book which would have, roughly speaking, about seven readers. Because much of this isn't much fun.

Actually, the tech part isn't so hard: This story is much more Gordon Gekko than Gordon, Flash. But you have to understand some of the tech side to make sense of the investment-opportunities side, so bear with me. I'm going to look at the tech stuff today, and the emerging marketplace tomorrow. It will be dense reading, but worth it; fasten your seat belts.

Fortunately for investors, it's not hard to find some companies that look like prospective winners amid this confusion, once we untangle the claims and geek-ology. So let's go.

The driving force here: There's a huge demand for faster Internet access. When Compaq (CPQ:NYSE), about to come to market near the end of last year with "ADSL-ready" PCs, conducted market research among potential customers, more than 70% said they'd happily pony up for a faster Net connection.

No wonder. Today's fastest analog modems are rated at 56Kbs, though connections are limited by law to 53Kbps. In practice, few PC users ever get anything like 53Kbps, because telco line quality and the likely presence of switches along the signal path from your home to the telco office -- switches which flip the signal from analog to digital, and back -- severely limit performance. Moreover, many PC users still use 33.6Kpbs and 28.8Kbps modems -- which also rarely deliver their full rated speed.

On average, U.S. Web users probably average 25Kbps-30Kbps connections to the Web from their homes. If that. Which leads, of course, to jokes like the "World Wide Wait."

Three technologies are competing to replace these analog devices as we move into the digital world: ISDN, xDSL and cable access to the Web. A fourth path, satellite access through pizza-pan-sized mini-dishes, is being pushed by Hughes Electronics' (GMH:NYSE) DirecPC, DirecDuo and USSB divisions, but has gone almost nowhere and has a dim future.

ISDN has been around for two decades, and thanks both to advances in other technologies and (mainly) the telcos' traditional and continuing stupidity, has never taken off ... and likely never will. ISDN lets you open two "pipes" along standard copper phone wiring, in a single phone line, each capable of carrying 64Kbps. One can be used for voice conversations while the other is used for data communications. Or you can "bond" the two into one 128Kbps data connection. ISDN comes close to delivering the speed claimed for it, but in the face of faster and cheaper alternatives, that only-slightly-better speed is no longer very interesting.

And the RBOCs -- the former "Regional Bell Operating Companies," such as Bell Atlantic (BEL:NYSE) and SBC Communications (SBC:NYSE), etc.) -- have wildly overpriced ISDN lines, usually charging a monthly flat fee plus a connect-time surcharge. It's easy to spend $100-$150 a month for mediocre ISDN service from most RBOCs.

--------------------------------------------------------------------------------
Needlessly high pricing arises from the core problem in buying faster Net access from the telcos: The RBOCs are desperate to protect their sky-high T1-line rate cards.
--------------------------------------------------------------------------------

That needlessly high pricing arises from the core problem in buying faster Net access from the telcos: The RBOCs are desperate to protect their sky-high T1-line rate cards. With a T1 priced at from $1,200 to $1,600 per month in most areas, plus a $100-$300 per month ISP charge to manage the line, plus an expensive router on your premises to make it all work, T1s produce huge revenues for the telcos. Anything -- ISDN included -- which threatens that gouge is anathema to the RBOCs.
It's been obvious for years that those high T1 prices couldn't last, but the retrograde thinking of RBOC managements across the country has led them to embrace the old and resist the new with all their might. RBOC execs talk a good game about the new, but with so much of their revenues wedded to the old, look askance at anything truly new. Like digital telephony. I've looked in vain for years for signs of a progressive RBOC CEO with the courage to truly embrace the digital revolution and lead his or her company into the digital age. That would have been a fabulously profitable move. But I might as well have been watching for Martians in the U.S. Senate. (Hmm....)

That's why today ISDN -- known by a wide range of insulting terms, principally "I Still Don't Need it" -- is toast.

The real contest is between the telcos and the cable companies, each with a favorite new technology, most of the needed infrastructure in place, and huge customer bases into which to sell fast access. And each with a huge hidden agenda.

Cable outfits are pushing Internet access over their existing cable connections. The telcos want instead to convert your residential phone line to xDSL. (The "DSL" part means "Digital Subscriber Line," and we use a leading x to indicate that there are actually several flavors of DSL -- ADSL, SDSL, HDSL, VDSL, etc. For home consumption, we're talking about ADSL, or Asymmetric DSL, in which the speed of transfers down to your PC is much faster than the speed of transfers up from your PC to the system, and thence to a Web site.)

Cable modem access is, and will likely remain, a better bet for most users. The cable companies typically charge about $40 per month for fast access, including ISP charges. (They take the place of your existing ISP, and typically offer some useless "value-added" ISP-like services. The fast access is great; the value-adds are a joke.)

While cable modems can in theory deliver up to 10Mbps, maybe even 30Mbps, to your home PC, in practice they're typically limited by your PC, your network card, and cable outfits' sloth to about a twelfth or fifteenth of that -- on average, typical speeds of 700Kbps-800Kbps. Moreover, you share the available bandwidth with some unknown number of neighbors with cable-Internet-access service, whom the cable company has thrown into what they call a node with you. If no one else is on, you're likely to get the full 700Kbps or so available in your neighborhood. If there are several other Netheads in you neighborhood, and all are using the Web at the same time, your access speed can slow to as little as 50Kbps or so -- conceivably, even less.

Hooking up cable access is easy. You need a standard 10Mbps or 10/100Mbps Ethernet card in your PC, plus a "cable modem." (Actually it's a router, but the cable industry is terrified that terms such as "router" will scare away the technophobes, so they misname the device a modem, which it manifestly is not, because in a digital world there's nothing to MOdulate or DEModulate.... In fairness, cable-system management may simply misunderstand what's in the "modem" box.) The best-known cable modem today is Motorola's (MOT:NYSE) CyberSurfr, but at the Consumer Electronics Show in Las Vegas earlier this month, Cisco (CSCO:Nasdaq) talked about its plans to be big in the cable-modem business, too.)

Cable companies typically provide the cable modem, folded into your monthly fee on a lease basis, and if necessary will usually install the needed network card in your PC for a nominal charge, as well. Initially, as cable companies scramble to grow their customer bases before the telcos come calling at your door, you'll see a lot of "Free Installation!" and "Free Network Card!" deals. Maybe even "Free Modem!" too, though I haven't yet seen that.

Cable access is very quick -- most pages appear almost instantly, and even longish downloads are rapid. For reference, TheStreet.com's home page comes down to my cable-modem-equipped PC in about two seconds. And I can download big files, such as Netscape (NSCP:Nasdaq) Navigator, in just three or four minutes.

Perhaps best of all, cable-modem access to the Web is (like a T1 connection) a continuous connection: You're always online, so you don't have that irritating 30-second series of screeches from your analog modem (properly: "negotiation") each time you connect. (Sound odd? You don't turn your cable-TV connection off, do you? You can always switch channels and the next one is right there, eh? Same with cable modems.)

In fact, cable access is close to nirvana for Web users. With time, as we get jaded and Web sites start to exploit broadband (read: fast) Web connectivity, we'll get antsy even at 600Kbps-800Kbps, and demand more. It'll be there; it's relatively easy and cheap for cable providers to crank-up system performance (for now they have no incentive to do so). (Yahoo! (YHOO:Nasdaq) already has a group assigned full-time to developing TurboYahoo, which will more fully exploit broadband connections. The company will start creeping TurboYahoo features onto its main Yahoo site slowly, as statistics show more and more of us have fast Net access from home.)

The real problem with buying your Net access from your cable company is...buying your Net access from your cable company. Many local cable outfits have a richly deserved reputation for unresponsive customer support, and you can probably expect that to carry over into their Net-access business. It wasn't an accident that Jim Carrey's dark 1996 movie was called The Cable Guy.

Both cable access and ADSL have big hidden gotchas.

On the cable side, it's security. Conceptually, your PC connected to a cable modem is actually a node on a network. That means that if, say, a neighborhood teen-age hacker lurks on a similarly cable-connected PC, he can with relative ease get into your PC, read your files, run programs, erase your hard disk. "Personal firewall" answers are on the way -- and for now, Windows users can greatly decrease their exposure just by turning off file sharing in Windows 95/98 -- but for now, cable-access users live in an insecure world. @Home (ATHM:Nasdaq) has already faced a security problem in the San Francisco area.

On the other side of the fence, the RBOCs are finally making Big Moves to establish ADSL service as the preferred fast-access choice for consumers. Since the first of the year we've had announcements such as the Bell Atlantic (BEL:NYSE) deal with America Online (AOL:NYSE) for a joint venture to deliver ADSL service for AOL customers in BEL's five-state area for $20 per month (on top of AOL's $21.95 monthly charge). We've seen SBC (SBC:NYSE), aka Southwestern Bell, talking about ramping-up ADSL availability in its service areas: Texas, Missouri, Kansas, Arkansas, Oklahoma. We've seen PacBell (SBC's satrapy in California) announce a rate cut of from $89-$329 to $49-$159 for its existing home ADSL service. And so on.

After years of punting and avoiding the issue, the RBOCs are finally awakening. At the same time, out of about 500,000 fast-access-to-the-home customers in the U.S. today, about 475,000 have cable modems; ADSL connections hold only about 5% of the market. ADSL obviously has a long way to go.

In theory, ADSL is a very appealing alternative to cable modems. It has a maximum download speed of about 8Mbs -- better than a T1 line -- and an upload speed of 758Kbps (more than enough, because you're usually just sending brief commands, such as a URL address, upstream). ADSL uses the existing copper wiring in your walls. It can easily deliver voice and data service on one line, so you don't have to pay for a second, separate, datacom line.

Unfortunately, few U.S. customers will see anything like that speed. Because the RBOC's are lining up behind an intentionally crippled flavor of ADSL called G.lite, which at its very best can deliver 1.5Mbs down to you, and 512Kbps upstream. G.lite ADSL has one huge advantage for the telcos: no "truck rolls." Standard ADSL requires a POTS ("plain old telephone service") splitter both at your home and also at the telco's office, as an interface to the analog system, to give you voice service. Which means the telco has to send a truck to your house to install that splitter -- about a five-minute, turn-four-screws job, The telcos hate truck-rolls, which they see as a huge, irritating and ultimately unnecessary expense. So they're backing G.lite ADSL to save the one-time cost of a truck-roll to your home.

A 1.5Mbs link to the Web -- about the real-world performance of an expensive T1 line -- is still darned fast, and beats today's typical cable-modem speeds. So what's wrong with G.lite ADSL from the customer's perspective?

Beyond the needlessly compromised performance, thanks to the telcos' traditional fear of undercutting their T1 revenues, it's the ridiculous pricing. Look closely at the RBOCs' new ADSL offers, and you'll find that they either cap service at 256Kbps or 384Kbps, or they're offering "tiered service," with 256Kbs for $50-$80, and faster service -- often capped at much less than full G.lite ADSL rates -- for $150-$175 or more.

In other words, the telcos' gouge continues. Less for more. Because they're scared of giving you a better deal. For users, the ADSL gotcha is fierce: ADSL only works if you're fairly close to the nearest telco switching office. The usual maximum distance is three to four miles as the cable flies. And if you're more than about three-quarters down that line, performance inescapably falls off: You may not get even the capped speed.

No matter what your local telco may tell you in those cutesy inserts with your monthly phone bill, remember that this contest is not a technology issue: It's a monopoly's pricing decision. (And you thought Microsoft (MSFT:Nasdaq) was heavy-handed!)

So who wins here? Which technology will prevail?

I think both will survive and prosper. This isn't going to be VHS/Betamax, where an inferior technology overwhelms and crushes a superior one. Despite the market machinations, we're going to see very large installed bases of both cable modems and ADSL modems over the next few years.



To: Jay Lowe who wrote (4308)1/16/1999 7:35:00 PM
From: Frank A. Coluccio  Read Replies (1) | Respond to of 29970
 
Jay, please clarify the following:

>>The DSL sub-nets are impoverished insofar as a specific, practical approach to load-leveling ... there are initiatives (concepts) but they aren't going to stop PacBell from melting their NAPs in '99-'00. <<

It seems to be contradictory to your argument favoring cable modems. It implies that the NAP would be overwhelmed by DSL carriers. In order for the NAP to be overwhelmed, or melted down, it would require that there be very little blocking in the DSL loop, which translates to high throughput. Is that what you are saying? Unless you mean that the absence of traffic due to low throughput in the DSL plant would cause congestion, possibly. But that doesn't make sense to me either. Explain, if you would.
---

We've not been covering all the bases here, IMO. In order for the NAP and peering nodes to be properly fed by serving ISPs, much less overwhelmed, their edge networks would first need to be bolstered. This is something that I've seen very little of discussed in these posts. And these upstream router ports and links between ISPs tiers and to the NAPs are not cheap. In some cases (depending on what's in place now) it will mean taking the next step up in router sizing, alogether.

In the case of ATHM, their planned backbone upgrade by T should handle a good part of this. Does anyone recall what that price tag is going to be for this?

In the case of the DSL ISPs, they will need to do something in kind [which is NOT the responsibility of the ILEC to finance, rather it falls on the individual ISPs to acquire new capacity from whomever]. The reason for this is that their current edge platforms will not support additional 50-fold levels of throughput.

Some of them cannot even support current levels, especially during peak periods, despite their using only a 28.8 framework.

Regardless of whose solutions are used, therefore, whether they be DSL or CM, the ISPs are going to have to bolster their router port sizing, and their transmission capacity to peering entities and to the NAPs.

The impact of these requirements on small ISPs can be pivotal to their future. Many of them are cash-strapped, and they will be even harder pressed if they do upgrade, and doomed if they don't.

Regards, Frank C.



To: Jay Lowe who wrote (4308)1/17/1999 3:43:00 AM
From: Jing Qian  Read Replies (3) | Respond to of 29970
 
>> Long-term -- optical and wireless slide in underneath and eventually dominate. Copper is scavenged for use in jewelry and kitchenware.<<

Optical to replace Cable modem? Is it a fact that the backbone net work which AT&T is building for @Home is completely a OC-48 based
DWDM network? OC-48 is in fact a fiber optical transmission protocol that can handle 2.5 Gbps bandwidth. @Home is well integrating fiber optic into their network. So fiber optical technology only benefit @Home, rather than a threat. If @Home doesn't have fiber optic in mind, how can they be confident enough to team up with RealNets to
lunch video streamming service?

I agree Wireless can become a threat to Cable in the long run. But
so far, there is nothing viable that can remotely compete with either cable or ADSL. I was one of the original developers of HNS's DirecPC product line and intimately familar with Satellite based Internet service. The max DirecPC can go is 400kbps, sure, a lot faster than 28k modem, but way slower than Cable. The biggest problem satellite access is the following:

1) Round trip delay. The signal has to be bounced back from 10 of thousands of miles to the Hughes satellite. Every download effort will
have to add that delay. But only the signal hits the dish, boom, it's fast. So you'll see the web page to be loaded all of a sudden after a 2 second or 4 second no show. This is bad for voice and video.

2) Inbound traffic is still thruough dialup modem at 28K. Satellite is only one way traffice, not too way.

So, we can pretty much rule out GEO Satellite access as a competition in any time. This kind of satellite is too faraway in the orbit to cause huge delay.

The only wireless threat to cable may be LOW ORBIT SATELLITE.
Teledesic may be the real enemy of ATHM in 5 years. But Teledesic hasn't lunched enough satellites yet. Plus they have to seamlessly hook up all the small satellite in low orbit. That will require years of
software development.

So, IMHO, I don't think Cable will have any real threat from now to 2005. Mean while, we should pay attention to the wireless front, finding out the wireless equivalent of ATHM.

Jing