Networking used to be a really interesting game. Hundreds of new companies whose names were filled with the letters X and Z offered all kinds of exciting technology. Level 2 switching gave way to Level 3, then 4, and soon we'll be up to Level 8.
Or maybe we'll be in the twilight zone, networking-wise, because there are strong indications that our old comfortable world is changing. All those X and Z companies are being replaced by three companies with the initials C, L and N. Why do Cisco Systems, Lucent Technologies and Nortel seem to be preparing to divide the world among themselves, and what will happen to the rest of us - vendors and buyers alike?
Behind all this vendor dancing is a basic change in the market. It's pricing. LAN switching is falling to less than $US100 per port. At that price, switching products can't be sold through a direct sales force and switch profit margins are tiny. The vendors that rely most on LAN sales are hit the hardest, so Cisco - whose router profits are enormous - is weathering the trend while 3Com and Cabletron Systems stumble a bit.
There is probably more stumbling ahead as prices continue to fall. LAN switching at 10Mbits per second is headed toward $50 per port, and that's going to mean new players. Who'd want to enter that kind of market? The companies that already sell in retail outlets. The kingpins of tomorrow's LAN market will be Intel, Compaq, Dell and Hewlett-Packard - the system and chip vendors. The profits these companies earn in their core markets and their consumer visibility growing out of those core markets will make them the big players in the LAN.
In the WAN, the number of network-dependent businesses is growing faster than end-user technology staff. Twenty years ago, almost all WAN buyer organisations had a high level of technical skills on staff. Today, fewer than one in three does. The result is a trend toward outsourcing enterprise networking to the carriers. That's why Lucent and Nortel are suddenly being cast as key players in a data market.
In a survey my firm conducted five years ago, a user responded to a question about Nortel by asking, "Is that that new start-up in Minnesota?" Today, that "new start-up" has bought Bay Networks, which was once seen as Cisco's archcompetitor. Why? Because the WAN's future belongs to the carrier virtual private network (VPN) service, not the private enterprise network, and Nortel has a strong position with carriers.
The sale of Bay, the biggest of the second-tier players, is a sure sign that the changes in the market are going to have an impact on the rest of the second-tier companies. 3Com, Cabletron, Newbridge and even Ascend are probably going to have to band together as a group or be bought. Besides the Big Three, the buyer candidates are a who's who of the carrier world, notably Siemens, Alcatel and Ericsson. Imagine a Siemens executive in a business suit arm-wrestling with a Cabletron guy, a la former Cabletron CEO Robert Levine - the mind boggles.
The buy-or-be-bought frenzy is affecting new market entrants as well. Three out of every 10 start-ups are nonviable right out of the box. They're being puffed up by clever merchandising to be sold to somebody, not to sell products. Even viable new market entrants may have to be acquired to sell their products, because carriers are conservative in committing their traffic to untried vendors.
Is this going to change networking? You bet. We're headed toward a marketplace in which giants stomp about and niche players scurry between their feet. Once in a while, a niche player will eat some high-protein market segment and get large enough to challenge the bigger players, but only once in a while. Hardware technology issues are going to become increasingly secondary to price issues.
But the really interesting stuff will happen when cheap switching dominates the LAN and VPNs provide businesses access to carrier services with no capital cost and no complicated projects. For example, we'll be able to set up a VPN for a two-hour collaborative conference. The focus will shift to software for applications in the network. That's already happened in the PC market. Read an issue of a PC publication and see how many software articles and ads there are, compared with PC articles and ads.
Networks serve businesses and users, but neither consumes the hardware. Applications that run on nets are what consumers and managers are buying. The focus on building an infrastructure to run those applications is logical when that structure isn't in place or isn't adequate. We're fixing the hardware problems of networking, and we're getting ready to go to the next level.
That brings us back to the Compaqs, Dells, Intels and HPs of the world. As prices fall and the base of users broadens, networking of computers increasingly will be the domain of consumer-level vendors. We've been tolerant, if a little condescending, about the offerings of these and other system players, and some (such as Digital) haven't made a big success of their networking attempts. Nevertheless, these guys have the right formula: They touch the user directly with their products, their names are recognised, they have retail channels to move products through and they sell enough network/system elements to earn a good profit from the whole package.
Maybe 10 years from now we'll be arguing over whether Compaq will buy Lucent, not the other way around.
Thomas Nolle is president of CIMI, a technology assessment firm in New Jersey, US. He can be reached at +1 (609) 753-0004 or firstname.lastname@example.org.