OUTLOOK: INFRASTRUCTURE

On a single day this past January, Cisco spent $567 million to buy not one, but two, virtual private network (VPN) equipment makers.

The signal couldn't be clearer: Virtual private networking is going to be the hottest infrastructure growth area this year. Through the acquisitions, Cisco revealed that it is no longer banking on tying VPN technology to its routers exclusively. Now Cisco has remote-access enterprise VPN gear from Altiga Networks and carrier equipment from Compatible Systems.

In the short term, VPN vendors will benefit from Cisco's moves. "Everybody's going to make some money in 2000," says Jeff Wilson, an analyst at Infonetics Research in San Jose. Makers of free-standing VPN devices will complete more sales because they won't face the powerful Cisco mantra that states free-standing gear is the wrong way to go, Wilson says. Look for vendors such as Check Point Software, Indus River, Sonic Wall and Watchguard to clean up, he adds.

The money is there. Enterprise users will plunk down nearly $300 million more this year for VPN equipment than they did last year, an increase of 65%, says IDC in Framingham, Mass. They'll use this equipment to build remote access and site-to-site VPNs for corporate employees only. With the momentum begun this year, by 2002 VPN spending will increase to more than $700 million per year, IDC says. Expect to see an uptick in VPN services this year as vendors ready for that splurge. These services will blend VPN technology into WANs. Likewise, VPNs and firewalls will become firmly tied together by the start of next year. And, VPNs will become integrated with public-key infrastructures for establishing user identities and determining encryption keys.

Vendors also expect that VPN capability in Microsoft's Windows 2000 will boost overall demand - if for no other reason than the client software will be so widely available.

Second to VPNs, this will be the year that Gigabit Ethernet and Layer 3 switches march into the enterprise. "Gigabit Ethernet will finally live up to its billing in 2000," says Lauri Vickers, an analyst with Cahners In-Stat Group in Scottsdale, Ariz. In both areas, expect to see Extreme Networks emerge as a bigger player.

A key new driver this year will be Gigabit Ethernet over copper, which started shipping at the end of 1999. Gigabit Ethernet over copper is attractive because it runs over existing Category 5 wiring and it costs one-third less per port than Gigabit Ethernet over fiber.

In 1999, Cisco led the Gigabit Ethernet pack with 39% of market revenue, In-Stat says. 3Com came in at a distant second, with 13% of revenue. That latter figure includes revenue 3Com earned from the Extreme Networks switches it resold.

Of course, 3Com has left the race and, as part of its exit strategy, is referring CoreBuilder customers to Extreme.

While its unclear how much Extreme will benefit from that move, it's relatively certain that the vendor will at least get looked at by new enterprise prospects. In fact, all smaller players, including Alteon WebSystems and Foundry Networks, should fare well from the Gigabit Ethernet boom. When it comes to Layer 3 switching, Cisco will continue to pull in the most sales, followed by Cabletron and then Extreme, according to In-Stat. With Cabletron's recent four-way split, continued growth in enterprise networking now falls under the unproven Enterasys Networks banner.

"The best way I can think of describing it is as a well-positioned wild card," Vickers says.

Here too, Extreme is positioned to pick up customers reluctant to go with the Cabletron spin-off.

In all, vendors sold $1.8 billion in Gigabit Ethernet switches last year and will sell $4.3 billion this year, In-Stat projects. For Layer 3 switches, sales were $2.6 billion in 1999 and are projected to be $5.3 billion in 2000. Meanwhile, Layer 2 switches still account for 79% of LAN switch purchases, In-Stat says.

The upshot: The drive for speed in the enterprise as well as the expanded use of VPNs will be key sources of revenue growth for network vendors this year.

--Tim Greene

3Com
On March 20, 3Com announced that it's exiting the enterprise network market. Toward that end, it will:

Lucent
On March 1, Lucent announced a restructuring. Per the plan:

Cabletron
On Feb. 10, Cabletron announced its breakup into four independent operating companies. Serving enterprise customers will be:

OUTLOOK: CARRIERS & ISPS

When the all-IP national broadband carrier Level 3 Communications burst onto the scene a couple of years ago, it dreamed of busting up the AT&T/MCI WorldCom/Sprint long-distance cartel. For nearly a year and a half, Level 3 executives from CEO James Crowe down banged the drums, beating out the message for enterprise users that its long-distance voice-over-IP service would smash per-minute tolls to bits.

But when the service was finally introduced last December, it caused barely a whimper among enterprise users. Why? Simple: If you run a corporate network, you can't buy it. Like almost everything else from Level 3, the service is for carriers.

Level 3's transformation from an enterprise network voice and data service provider in 1998 to a wholesale bandwidth vendor in 2000 is emblematic of what's happening in the carrier marketplace. More than 600 long-distance carriers operate in the U.S. today, yet many don't court enterprise customers. This trend will become the defining characteristic of the carrier and ISP marketplace by the end of 2000.

The pattern is already clear. While dozens of carriers start up every month in the U.S. and Western Europe, many are little more than shells - often started with junk bonds and headed by a few former MCI WorldCom guys. These carriers get into the action by renting everything from bandwidth to simple applications such as voice mail. The competitive local exchange carrier (CLEC) industry is booming, too, having grown 152% last year to $23.6 billion, according to New Paradigm Resources Group (NPRG) in Chicago. But many CLECs offer little more than alternative dial tone and T-1s - NPRG says 60% of CLEC revenue is voice-related, for all the image of CLECs as pioneers - and quite a few cater to small and midsize businesses.

In fact, in market after market, the raw statistics mask the fact that most carriers and ISPs do only a few things well: Concentration and core competencies remain the orders of the day. That's one of the reasons no one's sure which of the many application service providers (ASP) will meet with success this year.

Almost every major carrier has announced an ASP strategy or at least a partnership that purports to provide a strategy. Yet the market for pure applications hosting is growing relatively slowly (or only moderately) - from $1.1 billion in 1999 to an estimated $1.7 billion this year and $2.4 billion in 2001, according to the Yankee Group in Boston. Analysts say the market for real applications management belongs to a handful of specialized providers, particularly Corio, Interliant and USinternetworking. Much the same can be said for the Web-hosting business, dominated by ISPs that have branched out - such as MCI WorldCom's UUNET and Frontier's GlobalCenter - and a few independents like Exodus Communications.

AT&T's struggle to play in these new areas is particularly telling of how only a few will remain in the hosting and ASP businesses, for the long and short haul. One of the long-time major clients for AT&T's outsourcing unit, AT&T Solutions, is Merrill Lynch. So when Merrill Lynch decided to enter the online brokerage business last year, it looked at AT&T's plan to build 26 Web-hosting centers. Problem was, it was just that - a plan - and AT&T had barely begun building its New York hosting facility. So Merrill Lynch chose Exodus instead - and then smoothed out ruffled feathers by announcing that AT&T Solutions, Exodus and Merrill had entered into a three-way "partnership" for the firm's online venture.

It's the same story in the ASP space, where AT&T announced partnerships with no fewer than 17 server, application and operating system vendors before finally forging an alliance in February with the evident ASP leader, USi, to go to market. And until AT&T finishes building its Web-hosting centers - many of which are really converted central offices and aren't expected to come online until late this year or early next - we don't know whether AT&T will be a strong ASP contender.

Of course, only a few companies will suceed in busting up carrier markets at all. The outstanding example to date is Qwest Communications, which despite its public marketing is practically an "all things to all people" carrier: voice over circuit switching or IP to consumers or businesses, data over frame relay, ATM or Internet, and pure transport or applications hosting.

The reason more won't follow in Qwest's footsteps in 2000 is the tough ride the carrier had last year, and this year already promises more of the same. In 1999, it struggled against user reports of months-long circuit order delays, an acknowledged shortage of electronic equipment and slamming accusations from the Federal Communications Commission. This year, it's faced the resignation of Solomon Trujillo, CEO of merger partner US West, in open disagreement with Qwest CEO Joe Nacchio.

Of course, the regional Bell operating companies haven't given up the cause yet and will continue to fight hard in 2000. But so far, only New York allows the Bell there - Bell Atlantic - to sell long-distance voice and data services. By the end of the year, look for more than a half-dozen states to give the long-distance nod to their Bells. Then again, if Bell Atlantic's actions serve as the precedent, don't expect much competition to revolutionize the market. Bell Atlantic offers only a 10-cent-per-minute calling plan.

The upshot: Setting up a national carrier that has a complete business-user, customer-service structure will continue to be hard work.

-- David Rohde

OUTLOOK: ENTERPRISE APPLICATIONS

It wasn't so long ago that PeopleSoft's new plan to offer a service that lets customers rent applications, rather than license them outright, would have sounded like corporate suicide. But, as they say, the Web changes everything.

"Our upcoming PeopleSoft release 8.0 will have a 100% HTML user interface. We've done away with the Microsoft Windows client," says Andy Allbricten, vice president for PeopleSoft's recently announced "eCenter." The center hosts the company's applications and lets business customers access them, for a monthly fee, from any Web browser.

The Web has been driving the software industry to transform enterprise applications into browser-accessible services.

This year, all the major business software vendors will adopt that form. And in three to five years, software-as-a-service will become an accepted way to outfit an enterprise with applications.

Application providers expect to make a killing. PeopleSoft executives predict that within three years eCenter hosting will account for 50% of license revenue. Oracle Chairman Larry Ellison predicts a majority of license revenue would come from this ASP model within five years.

Overall, IDC projects the market for software applications and attendant services will hit $70 billion this year. By 2004, the market will soar to more than $300 billion. About $150 billion of the larger figure is for the software itself, the rest for consulting, maintenance and similar services.

But enterprises won't have to wait that long to reap the benefits. In 2000, because of the Web browser and the move to server-based computing, business applications will become more affordable, faster to deploy and easier to use than ever before. They will also become more "interrelated."

The vendors offering the most value will be those with an architecture that lets service providers incorporate other applications, and juggle data and business processes across different organizations linked via an extranet.

Throughout this year, enterprise users will see more attempts by independent software vendors and application service providers (ASP), in partnership, to create interrelated applications. "The vendors are tying into collaboration and workflow engines," says Dennis Byron, IDC's research director for enterprise applications. "Some industries, like architecture/construction, need a lot of collaboration for projects. With an ASP model, they don't have to create the infrastructure for this themselves."

Also in 2000, as these hosted services become more widespread, wireless devices will be created to access them, says Mark Gorenberg, partner with Hummer Winblad Venture Partners, a San Francisco venture fund specializing in software. In many industries, he says, business managers are starting to realize that wireless access to Web-based business data creates real-time interactivity. "There's a growing expectation that people can and should respond at once on mission-critical applications running on extranets and intranets," he says.

That realization will hit home this year with more companies writing applications that enable handheld devices to read and update Web-accessible data and line-of-business applications.

The upshot: Applications hosting will lower costs for enterprises while spurring demand for interrelated applications and wireless data access. -- John Cox

Contact Senior Editor Tim Greene

Contact Senior Editor David Rohde

Contact Senior Editor John Cox

Cisco boosts LAN management
Network World, 04/04/00.

Cisco preps more voice/data convergence wares
Network World, 03/28/00.

MCI wants Sprint for its wireless data service
Network World, 04/07/00.

MCI WorldCom pushes U.S., WTO to open Mexican markets
Network World, 03/31/00.

PeopleSoft details ASP plans
Network World, 03/08/00.

PeopleSoft CEO looks ahead to raft of new products
Network World, 02/03/00.

Send this article to a colleague