Where has all the cool net gear gone?

By Jeff Caruso
Network World, 08/02/99

Innovation has slowed to a crawl in the enterprise network equipment market.

For proof, just check out the kinds of new companies being funded by the top venture capital firms. Institutional Venture Partners, for instance, invested millions of dollars in four enterprise network equipment makers in 1996, but the Menlo Park, Calif., firm has invested in none since.

This is not to say that innovation has entirely disappeared. It's more that vendors and investors expect enterprise customers to start spending more of their money on managed network services and other offerings - and less on customer premises equipment.

Most of the money earmarked for high-tech companies these days is flowing into emerging service providers, their equipment suppliers and Internet software companies - not makers of enterprise network gear. According to the quarterly Network World/ PricewaterhouseCoopers Venture Capital Survey, nearly 75% of the more than $2 billion in venture capital spent during the first quarter went toward Internet companies (NW, May 24, page 1).

"It's not even fun to play in the enterprise market anymore," says Tony Sun, managing general partner at New York-based Venrock Associates, which now largely invests in companies that cater to service providers and small businesses. "Forget the products. It's all about distribution."

Customers are also noticing the trend. "I do see the emphasis from most of the product providers going into the carrier market these days," says Mike Ackermann, manager of network planning and design at Blue Cross/Blue Shield in Detroit. His company is already being pitched equipment usually aimed at carriers, as it considers deploying Nortel Networks' Passport switches. Ackermann says the switches are rock-solid and reliable, but more complex to support than typical enterprise gear.

Why this ambivalence toward the enterprise?

One reason is that Cisco's dominance of the enterprise market leaves little room for newcomers, observers say. And with the continuing commoditization of enterprise network hardware, Cisco and others with established distribution channels have a decided edge over start-ups.

Even leading network equipment makers are investing a smaller share of their research and development dollars on enterprise gear. Cisco, for example, now splits its R&D spending about evenly between enterprise and service provider technology - this from a company whose service provider business unit is only about a year old.

It also seems that there are few ideas being born that are so revolutionary they require new companies to promote them. Emerging technologies such as 10G bit/sec Ethernet, for instance, will probably just be added to existing products rather than form the basis of new companies.

The new breed of equipment makers - companies such as Juniper Networks and Atmosphere Networks - is attracted to the service provider equipment market, which market research firm International Data Corp. says is growing three times as fast as the enterprise network equipment market.

Many of the established network equipment firms, such as Cisco, Nortel and Cabletron, are actually playing both sides of the fence by catering to enterprise net customers and service providers. They are looking to deliver gear for building "end-to-end" networks that stretch from customer sites into the core of service provider networks. Some of their new products will be used for managed services, in which service providers locate equipment on customer sites, then monitor the gear for the customer.

Meanwhile, the last of the core enterprise network equipment start-ups are going public. Extreme Networks had its initial public offering in April, while Foundry Networks and Alteon WebSystems filed to go public earlier this summer.

But industry watchers don't think the companies will stay independent for long. Even FORE and Xylan - both of which successfully sold enterprise network gear for much of the 1990s despite competition from larger vendors - have been swallowed up in recent months by traditional telecom equipment makers.

Such acquisitions could further stifle initiative because telecom vendors are slower to move than smaller, nimbler companies, says Stan Schatt, research director at Giga Information Group in Boston. "We're in kind of a lull in the enterprise campus network" in terms of innovation, adds Paul Zagaeski, senior industry analyst at Giga.

For example, the demand just doesn't appear to be there for advanced voice and data convergence products. This is partly because it's so inexpensive to just throw bandwidth at surging network traffic loads. The lack of quality-of-service standards is also delaying voice/data convergence projects, observers say.

Industry watchers see opportunities for innovation at the fringes of the enterprise network, rather than in core switching and routing functions. Storage-area networks are one emerging area. Switches that balance loads among Web servers are another, although these are used primarily by Web-hosting service providers. Edge devices that provide for virtual private networks are also gaining in popularity and have become something of a start-up hot spot.

But the age of the spunky start-up going after parts of the network ruled by Cisco seems to be coming to a close.

"It's hard to go after that market because the distribution channel is so splintered," says Michael Speyer, an analyst with The Yankee Group. "How do you market to 10,000 value-added resellers if you don't have a huge war chest?" o