The biggest questions raised by this week's announcement of an enterprise access point from D-Link remain unanswered.

Since wireless networks have been suggested for offices, the architecture they should adopt has been hotly debated (see our article from last year on the Wi-Fi architecture wars).

New entrants to the office wireless market, such as Trapeze (see Trapeze Mobility System review, Symbol (see Symbol Wireless Switch System review), Airespace and Aruba have promoted the idea that access points designed for offices should operate only under the control of a central switch.

These enterprise access points are sometimes called "dumb", sometimes "thin" and sometimes "lightweight" - but all these terms are misnomers. The points are obedient to central control by a switch and, until now, have been made by that vendor. They operate as part of a system. So far, however, enterprise access points have been proprietary.

The main objector to lightweight access points has been Cisco - and its own wireless management system is equally, if not more, proprietary.

This has been a major reason for IT managers to hesitate over enterprise Wi-Fi. Buy a switch, and you pay a lot of money for proprietary Wi-Fi ports and tie yourself to buying access points from one vendor - a vendor that is funded by venture capital and operating in a fluid technological environment.

So, on the face of it, the arrival of D-Link should make a big difference.

Could D-Link add stability to the mix?
As a major player in standalone Soho Wi-Fi, and low-cost network hardware generally, D-Link is not going away. The access point it proposes to sell will comply with a network protocol (LWAPP) intended to standardise the way access points talk to wireless switches.

It could be a lot safer to buy into a wireless switching architecture, if you know that - even of the switch vendor goes out of business - you can still get access points from D-Link.

However, the standard doesn't quite exist yet and nor does D-Link's enterprise access point. The IETF's proposed CAPWAP protocol will be developed from Airespace's LWAPP protocol.

Until the standard is completed, any LWAPP access point can only communicate with Airespace's switches. This would be great news for Airespace, as it would have pretty much the only wireless switch which works with different vendors' lightweight access points.

Of course, we do already have systems from people like Vernier (as sold by HP - see HP700wl review) which operate with, and provide a measure of control over, access points (thin or fat) from multiple vendors.

What is D-Link really up to?
The multi-vendor access point system only exists when D-Link delivers its products, and our guess is it has no intention of doing so until LWAPP is closer to becoming a standard. If this access point is sold as an LWAPP AP, adding to existing LWAPP based systems, then it won't have much of a market until LWAPP becomes CAPWAP. Until that day, the only LWAPP systems will be that of Airespace and that makes for a small market in D-Link's terms.

Unless of course, the AP is not just an extension. D-Link may plan to sell it as a standalone AP, with LWAPP added simply as a badge of enterprise credibility, a reassurance to network managers that it will extend to fit into enterprise systems if, and when, LWAPP becomes a successful standard.

Alternatively, D-Link (which has a perfectly good switch heritage) may decide to bring out its own LWAPP switch. Before Christmas it was hinting that it might have an enterprise switch in the pipeline. It may be that those hints simply referred to this access point, but we think that it would be strange for D-Link to address (and commoditise) the access point, without going for a piece of the higher-revenue action in switches.

What price the enterprise access point?
One thing that users will look for in an eventual enterprise access point will be price. They want them cheap and reliable, but that still gives D-Link plenty of room to make a premium product that sells for more than a Soho AP.

"Enterprise access points sell for £300 to £4,000, three or four times the price of consumer products," said analyst Richard Webb, of Infonetics Research. "The demand for enterprise WLANs continues to grow at a feverish pitch, with worldwide demand surpassing $300 million in Q3 2003 and expected to reach $2 billion by 2006."

A slice of this pie would be most welcome for D-Link and other wireless vendors currently selling to the Soho market. Some have predicted that their arrival will lead to the $50 enterprise access point, with hooks to manage RF performance as well as CAPWAP. D-Link will be hoping for a little more than that, but will be prepared to lower the cost of enterprise APs somewhat.

Airespace has sponsored a white paper on the evolution of the enterprise access point which is worth reading.