By Paul McNamara, Chris Nerney and David Rohde
Network World, 9/28/98

Running an enterprise network is tough enough without the dull roar that accompanies this business - the hype from vendors, analysts and the press about the Next Big Thing in networking.

But you can put the energy to work for you. How? By learning the Cycle of Hype.

What is the Cycle of Hype? It is a predictable pattern of events in a technology's life cycle that can help you determine whether a particular technology is really ripe for installation. Years of experience show that on the way up the hype curve, you may be able to determine that a technology is not ready for prime time. Look for any of the following five characteristics:

1. The technology was driven by something other than user need, such as the desire to crush a hated competitor.

2. The technology seems to solve a relatively unimportant problem or one that's not relevant to enterprise networks.

3. The technology is expensive and requires forklift upgrades.

4. It's not scalable to the enterprise - at least not yet.

5. It's just too complicated for anyone to understand.

Look for these traits to determine when everyone is writing off a trend: Is the number of programmers working on that particular platform growing? If it's a carrier service, are numerous carriers continuing to expand the number of places where a particular service is being made available, indicating they garnered enough takers in their first markets?

The Cycle of Hype can be seen in hardware, software and carrier markets. Let's follow the Cycle of Hype as it influences some of this decade's most publicized companies and technologies.

So why did we need network computers?

In place of the Windows-based PC, Ellison and McNealy envisioned the network computer (NC): a desktop machine shorn of its hard drive and diskette drive but packed with memory that would retrieve data and applications from a network of servers. This would all be done for a desktop price of $500, we were told then, at a time when PCs were typically selling for three to five times as much.

Oh yes, and under this scenario, Microsoft was to get knocked down a peg or two. What's not to like? Nothing, of course. But the NC revolution promised by all that hype simply has yet to materialize. One reason is that the price of a PC has plummeted beyond anyone's wildest expectations. But that's not the only reason.

"It was a concept that was driven not by user demand, but by the hatred of Larry Ellison, Netscape and IBM of Microsoft," argues Bill Laberis, a former editor in chief of Computerworld who now runs a consulting company.

The dramatic potential of that titanic confrontation, rather than its technological import, powered much of the hype around NCs, particularly from the media, according to Laberis.

"The press ate up this NC concept, completely ignoring the most overwhelming fact of life in the corporate environment today, and that is the presence of Windows and the whole Windows environment," he says. Today Ellison and Oracle are in full revisionism mode concerning the NC, Sun is doing what it can to drum up excitement for its tardy JavaStation, and other NC vendors, while remaining in the hunt, look rather the worse for wear.

However, before counting the NC out on the canvas, it's worth noting that the Cycle of Hype does allow for comebacks. Internet pundit Bob Metcalfe, who invented Ethernet in 1973 and founded 3Com six years later, believes that NCs still have life in them.

"The NC will be off-stage for a while, and then reemerge with a new name, something like network appliance," Metcalfe speculates. "And everyone will believe that Larry [Ellison] was wrong, when he's right."

Push's 15 minutes of fame

As 1997 began, push technology and the Java programming language arguably were riding the largest hype waves in the industry. Analysts and pundits were rhapsodizing about how push literally would redefine how we use the Internet.

By the fall, however, the industry's infatuation with push was over, soured by a growing awareness of the technology's limitations and a standards squabble between Microsoft and Netscape.

Rather than being a paradigm-shifter or a huge new market, push in the end became - for the vast majority of users - nothing more than a free browser feature.

"Push is a fact. It exists," says John Rymer, president of Upstream Consulting of Emeryville, Calif. "But it has become just a part of Microsoft's product line."

Push evolved in direct response to user dissatisfaction with the limitations of Internet browsers and search engines. The premise behind push was simple: Instead of "surfing" endlessly through cyberspace for the information they need, users could have tailored information transmitted, or pushed, to their desktops via categorized channels.

PointCast Communications of Cupertino, Calif., was the first vendor on the market with a push product, launching its PointCast Network in February 1996. The free software overrode PC screensavers, displaying regularly updated news and information from a variety of sources.

It soon became a favorite download of corporate workers nationwide. They were intrigued by the notion of having the latest sports scores and stock prices awaiting them on their PC monitors when they came back from lunch - not exactly mission-critical enterprise applications.

Also in February 1996, ex-Sun product manager Kim Polese started Marimba, Inc. Rather than push content, Polese said, Marimba software would push enterprise applications to corporate workers.

Through most of 1996, the two companies generated almost all of the hype surrounding push. In the meantime, dozens of other start-ups clambered aboard the push bandwagon, hoping to make a killing in a hot market.

The push buzz grew louder in November 1996 at Comdex in Las Vegas. Push start-ups were among the stars of the show, and Netscape used the opportunity to unveil plans for incorporating the technology into its market-leading browser.

Within a few months, though, the push backlash began. Many users complained that their bandwidth-hogging push channels were causing their PCs to crash. For others, the novelty simply wore off, and information overload kicked in.

Even worse for the numerous push start-ups was the entry of Microsoft and Netscape into the game. By announcing intentions to incorporate push into their browsers, the two giants reduced the rest of the field to fighting over corporate niche markets.

They also began fighting for control over push standards, a battle that made many users leery of betting money on the new technology.

By the end of 1997, most push vendors had gone out of business or were bought. And those that survived changed their marketing strategies: Marimba now calls itself "a pioneer in application distribution and management."

And the former star of push technology, PointCast, continues to post huge losses and suffer dramatic customer turnover. A planned initial public offering (IPO) was canceled last July, despite a highly favorable climate for Internet-related stock offerings.

Hope, drama and Java

Perhaps more than Java's promise of "write-once, run-anywhere" application development, Sun's programming language owes its popularity, or "mindshare," to its role as a potential giant-killer.

"Java represented a strategic technology that could undermine Microsoft's control over the desktop environment," says Ron Rappaport, an analyst with Zona Research of Redwood City, Calif.

This meant Java could tap into the emotions of millions of Microsoft-haters in the overlapping Internet, network and computing universes.

Sun and other Microsoft competitors, notably Netscape, IBM and Oracle, played this card to the max. Sun CEO McNealy and Oracle CEO Ellison, in particular, began barnstorming around the trade-show circuit in 1995, whipping anti-Microsoft zealots into a frenzy by vowing to break the software maker's stranglehold over the desktop.

The trade and business press, always on the lookout for great drama, eagerly played along, casting the looming conflict as the latest Battle for the Soul of Computing. But to sustain interest for more than a few months, Java had to be more than the next contender.

"You also need a value proposition, and in Java's case, it's an Internet-oriented development language that for many developers is easy to use," Rappaport says.

Through 1996 and 1997, Java momentum and interest grew as the programming language continued to develop. Sun continued its tireless promotion of its franchise, signing up more licensees, establishing Java purity tests and attempting to gain control of the Java standards process.

Kleiner Perkins Caufield & Byers, the most prestigious venture capital firm in Silicon Valley, even started a $100 million Java Fund in August 1996 to finance start-ups with product plans based at least in part on the programming language.

By mid-1997, however, Java momentum had slowed, the victim in large part of extravagant promises by Sun that did not correspond with reality. Java performance was sluggish, it did not scale well and it was not particularly portable.

Programmers complained that Java applications were painfully slow compared to those written in C++, and the phrase "write-once, debug-everywhere" became a popular joke, as developers experienced numerous problems porting applications across different platforms.

To add insult to injury, Microsoft began developing its own line of Java software - development tools and Java Virtual Machines - that performed better than those being cranked out by the Java alliance.

Further, numerous product delays were causing corporate buyers to hold off on any commitments to the programming language.

The release by Microsoft in September 1997 of a Java-incompatible browser triggered a licensing lawsuit by Sun and accelerated a backlash against Java. Press coverage of Java began focusing less on whether Java would become truly cross-platform and more on how many versions of Java ultimately could become available. Yet, despite increasing skepticism about Java's potential, the programming language continues to make steady, incremental progress. The number of Java programmers continues to grow, and surveys show that Java is beginning to make serious inroads into the enterprise, particularly on the server side.

Ironically, these enterprise gains are being made even as the Java alliance has somewhat muted its pitch to the corporate IT market. These days you find Sun's McNealy talking more about how Java will revolutionize embedded consumer devices such as telephones and home appliances.

Bottom line: Java appears to be a technology destined to outlive its hype, even if it never makes its way into your enterprise applications.

Netscape's fall from grace

Back then, the Web was a brand-new, wide-eyed world, and Netscape Navigator was the be-all and end-all of Internet travel guides. At an age when most people are worrying about repaying student loans, Netscape co-founder Marc Andreessen was helping to launch that year's most lucrative IPO. His baby-faced mug was plastered on covers of mainstream newsmagazines. Netscape's stock quadrupled to almost $90 a share before Santa even had a chance to drop in on Mountain View, Calif.

In short, Netscape could do no wrong.

That is until Microsoft woke up from its slumber and began peppering the planet with Internet Explorer.

While the rest may not be exactly history, at least not yet, Netscape has struggled mightily to live up to those lofty expectations established during those heady times. Irrespective of what happens with the company's latest attempts to reinvent itself or Washington, D.C.'s legal offensive against Microsoft, there is little chance Netscape will ever again recapture that early glow of youth it once enjoyed.

"Netscape reached too far, too fast, with not that much in the tank to begin with," Laberis, the former editor in chief of Computerworld, says. "It had a browser, which somebody else was willing to give away for nothing."

The Netscape experience demonstrates the double-edged nature of stock-driven hype, according to Laberis.

"Netscape has been duly punished by the stock market for over-reaching," he says. "That's the same stock market, by the way, that gave the company its inflated market valuation in the first place."

However, those who are already writing Netscape's corporate obituary are doing so prematurely, according to one long-time industry watcher.

"Netscape is one of those companies that came and went and will return with little help from the press," says Internet pundit Metcalfe. "The Internet allowed them to go over our heads directly to customers."

Netscape's revenue grew 54% in 1997, showing that someone is buying the company's products, even as most industry analysts and the trade press are writing them off. Moreover, while the company's once-dominant share of the browser market has been eaten away by Microsoft, Navigator retains the No. 1 spot, and its adherents have been buoyed by Netscape's decision to throw open the source code for its Communicator suite. Restarting a fresh wave of Netscape hype undoubtedly would prove more difficult moving forward, says one expert, no matter how often the company attempts to reposition itself.

"Netscape has been very unsuccessful at interacting with the trades," contends David Strom, former trade press editor. "They don't understand how to do PR."

We didn't ask Strom, but perhaps he means this kind of thing: Netscape failed to respond to repeated interview requests regarding this story, including the written questions that were submitted.

Frame relay or ATM? Watch your timing

Hype has played a key role in frame relay for one specific reason. For the market to really take off, it had to drag in a particular vendor - AT&T. In the early and mid-1990s, and even today, AT&T had the lion's share of the nation's private-line users who were - and are - the prime candidates for migration to frame relay.

Although AT&T technically announced frame relay service in 1993, it really only got rolling in 1995. It wasn't that no one wanted the service. Quite the contrary, in 1994 AT&T received lots of orders for frame relay but couldn't fill them because it hadn't installed enough ports in its network.

"The first decision users always make is to buy AT&T or not to buy AT&T," said one close observer of the industry. Until AT&T was convinced that it had more to gain than lose by pushing frame relay, the market was bound to struggle. The consistent drumbeat of education from competitors, press and analysts helped convince AT&T to forge ahead with the service to the point where its frame relay network crash last April made headlines around the world.

Where hype has proven to be a problem for frame relay has been in making the service something more than just a LAN-to-LAN interconnection protocol for non-mission-critical traffic running over IP or IPX.

For example, the key issue of SNA over frame relay has followed a traditional hype curve. In October 1995, Sprint announced the first substantial managed SNA-over-frame relay services, with a selection of frame relay access devices (FRAD) that the carrier would configure, install and maintain.

The problem was that most of the SNA users were on AT&T multidrop private lines, so the SNA-over-frame market didn't mean much until AT&T explicitly supported it. The company finally did so in November 1996, with options to place FRADs on the customer premise or in the central office. Even then, the effort seemed to be halfhearted, and earlier this year AT&T pulled the plug on the Central Office FRAD option (NW, July 27, page 8).

"It sounded like a winner, but several things contributed to Central Office FRAD's demise," says Tom Brophy, president of NetPlus, who was responsible for the development of the service while he worked for AT&T. "AT&T never actively marketed the solution or compensated its sales force accordingly because it was never generally available." He blames problems with AT&T's equipment supplier for AT&T's lack of confidence in the service.

Problems in figuring out which specific applications work on which networks have also been behind the yo-yo hype cycle of frame relay's sister service, ATM.

In the early to mid-1990s, ATM was touted as the technology most likely to run corporate networks over the next decade because it could handle voice, data and video applications with the appropriate quality of service.

The problem is that multimedia needs never were as important to enterprises as was the need to simply acquire bandwidth for data. Paradoxically, that need for bandwidth helps explain why ATM may be as under-hyped today as it was over-hyped yesterday.

All of the top long-distance carriers report that their ATM services are growing rapidly in 1998 for a simple reason: ATM supports T-3 data links, whereas frame relay usually maxes out at T-1. Builders of private corporate networks agree. "The adoption rate of ATM is probably higher than frame relay now," says Mark Kaplan, director of North American marketing for Newbridge Networks.

Lesson: If you hear that a particular technology is the Next Big Thing because it solves a problem you don't particularly care about, take it with a grain of salt. But make a mental note to check back in two years. The same technology could end up solving the problem you may care most deeply about - getting more bandwidth at cheaper prices in both the LAN and WAN.