It’s hard to remember the heady days of the 1990s, when network stocks were flying high, IPOs showed enormous gains in their very first hours and the financial future of networking seemed rosy. Networking is sure not looking like a high-flyer anymore. Will it come back?
Technology in general has been recovering from near-historical lows. During the worst of the post-bubble collapse (2003), IT spending as a percentage of the gross domestic product hit the lowest point since the post-World War II period. In the history of the Information Age, there has never been a period when spending stayed this low for more than a couple of years. That alone would seem to give hope for increased IT spending and maybe a couple of extra bucks for network equipment.
Then there’s the whole service-oriented architecture (SOA) thing. BusinessWeek has heralded SOA as the most significant software trend in a decade. The idea of SOA is to create flexible application relationships that let users build applications similar to the way that Web pages are built today. This means assembling pieces hosted on servers that might be in the data center, a partner’s server farm or even a service provider’s grid computing center. It’s hard to see how all this assembling could be done without generating network traffic, but you don’t hear much about the impact of SOA on networks.
One problem is that SOA might actually be undermining networking’s mind share with senior management. SOA is an application revolution, not a network revolution. It’s true that SOA might change the way that we do networking, and it will certainly change the demands placed on networks, but right now the focus of management is on the application and software. Companies that provide applications, such as IBM or SAP, are getting appointments with enterprise CIOs; network vendors are having a harder time. Cisco’s John Chambers talked about the need for a higher level of engagement last December; the question is whether anything can be done about it.
A second problem is commoditization of network equipment. There are just so many features that can be added to a LAN switch or router before a buyer simply doesn’t see the point anymore. LAN switching suffered a price collapse in the early 1990s because of this feature overload. Players such as the 3Com-Huawei joint venture, capable of operating at lower profit margins per sale, can hurt the profits of the major vendors when buyers increasingly want high speed and low cost. It doesn’t help that everyone’s chasing the small and midsize business market, either. This market isn’t as likely to value features, is too small for direct sales and is significantly more price-sensitive.
It seems likely that the carrier segment will have to provide the strength in the network market at least for the next couple of years. Consumer broadband and wireless growth are promising to drive up spending, and the steady upward trends in carrier capital expenditures have surprised Wall Street. It seems likely that content networking, such as IPTV, will be a major driver of spending, but will that spending be focused more on content servers and software than equipment? Is IT going to threaten network equipment everywhere?
It seems clear that the only answer to networking’s prayers for better performance is to become more involved in SOA and content trends. That doesn’t mean providing fat pipes at lower unit cost; it means finding network features that are specifically linked to SOA and content. Some vendors have made strides in this direction. Alcatel has IPTV-specific features in its DSL products, and Cisco has announced Application Oriented Networking, which in Europe is being billed as a way to make networks aware of SOA. But neither approach seems to be getting much promotion by its vendor, or much publicity and recognition.
Could it be that we’ve lost our excitement? Is the problem with networking now as simple as that the promoters don’t believe anymore? Is networking strategic or just plumbing? If we can’t prove it’s the former, it will surely become the latter.
Nolle is president of CIMI Corp., a technology assessment firm in Voorhees, N.J. He can be reached at (856) 753-0004 or tnolle@cimicorp.com.
