So, 802.11g and WPA security have both been approved, and wireless LANs on the 2.4GHz band have just got faster. 3Com has announced an 802.11g enterprise Wi-Fi product, and Cisco has launched a strategy for Wi-Fi.
All this is a signal for a bigger take-up of Wi-Fi in the enterprise. And that would be no bad thing, if the murmurings about the hype over public Wi-Fi services turn out to be true.
Some analysts are predicting that the Wi-Fi hype will lead to a dot-com style crash amongst public wireless ISPs, though I reckon that the hype is still manageable and the fundamentals are good.
So, if Wi-Fi service providers don't lay a golden egg, vendors will want to sell some more office-based Wi-Fi networks. Of course, it is quite possible that the market for office-based Wi-Fi is being hyped just as much as the service providers. To hear some people talk, you would think that the old wired office network has had its day, and that is certainly not the case.
Whatever the different markets for Wi-Fi do over the next few years, it is certain that right now there is a market for optimistic market research. What turns that research into a bubble is when too many people start to believe it, and forget that market research is usually commissioned by those with a stake in the market who want a bright future for people to invest in.
So, research aside, what is going to make enterprise Wi-Fi start to hum? WPA may be more significant technically, but 802.11g is getting publicity because speed sells. This is the case even though speed has not actually been a barrier to enterprise Wi-Fi (although users do share the bandwidth at an access point, so extra bandwidth means more users).
802.11g's arrival in the office will be powered by its previous success in consumer products. Whatever vendors in both spaces may say, the enterprise Wi-Fi market has a lot of synergy with the consumer market. We all know that products have to be different for the two spaces (if only to justify higher prices for business customers), but there's no denying one thing: the fact that there is a consumer market for this stuff makes it easier for the vendors to produce enterprise versions.
For instance, vendors have been selling "802.11g" products for at least six months to consumers. The standard has been pretty solid for a while, and perfectly ready for sales to the consumer market. While these products don't conform to a final standard, that doesn't matter to consumers, who usually have standalone networks with one access point. Users don't expect to be able to roam with their home Wi-Fi devices into their friends' houses.
Businesses, on the other hand, do need a full standard before they adopt a technology. They have had to wait, for .11g. However, the experience vendors have gained in selling 802.11g to consumers means they don't have to spend a long time preparing 802.11g products for businesses; they just need to tweak consumer-grade products to make them business grade.
So, it is no surprise to see 3Com launching a product almost immediately. Others are moving a bit slower, with Intel not expected to bring out its.11g products till the end of 2003. Symbol is being even more cautious, warning that early 802.11g products may have drawbacks, and won't have a .11g card till next year.
Symbol's delay may be surprising, since it was ahead of the other enterprise Wi-Fi vendors last September when it launched its wireless switch -- a specialised switch that sits in your wiring closet, handling all the wireless LANs that might be attached to your corporate network (and rooting out the ones that shouldn't be there).
As I've said before, the wireless switch has become this year's essential enterprise networking product. Symbol's is taking off well, apparently: networks based on it now make up 40 percent of the company's wireless LAN revenue -- a much faster take-off than Symbol expected.
So what problems does Symbol see with 802.11g? The company says the products will have trouble when working with .11b clients, including poor roaming and lower battery life. "Backward compatibility is very important to us because we believe that 'b' clients are going to be around for many years, probably another five years," says Gary Singh, senior director of worldwide marketing at Symbol. "The efficiencies that you get from power management in 'b' technology, don't exist in 'g' and 'a' today."
Fair enough: that's a testable statement, so I'd advise you to check roaming and battery life on any .11g product before you buy in any big numbers. Symbol is not the only one to give 802.11g a miss: Trapeze is concentrating on .11a, for the stated reason that it is better. It supports more channels and has less interference.
Another point is that the .11g products coming at the end of the year will include combined abg products. "At this point, the alphabet soup will go away," says analyst Richard Webb of Infonetics. Confusion with .11a is indeed an alphabet soup issue, not a technical one. Although .11g's 54Mbit/s speed spec is the same as .11a, neither Singh nor Webb expect a battle between the two. 802.11a is on 5GHz so it will be the standard of choice where interference with other 2.4GHz radio, such as Bluetooth is an issue.
But what of the Cisco launch? At a recent briefing, Symbol spent most of its time picking holes in Cisco's wireless strategy, instead of comparing its products with the other wireless switches that have come along since last September. This is because of Cisco's dominance. A Cisco launch into any networking area, usually late, and always at a higher price, is always seen as legitimising that area. And it sparks off a time of clarification, when other people figure out the distinguishing features of their own technologies.
Cisco's Structured Wireless Aware Network is no exception. It does not include a specialised wireless switch but bundles together other Cisco products and manages the wireless traffic on the existing Catalyst switches that vast numbers of enterprise customers already have. That may sound like a good idea, but Symbol's Singh dismisses it: "they had to come out with something," he says, "but it's really nothing but a product-bundling approach." Add up all the additional things you need from Cisco, and it comes to ten-times the cost of a switch-based Wi-Fi solution from a vendor such as Symbol, he says.
Other vendors are going to divide up according to how their wireless switches are architected. Do they aim to make you replace the switches in your wiring closet? Not surprisingly, switch vendors like Foundry seem to have solutions in this class. Or do they make an overlay alongside what you have already, saving existing investment? Symbol, Trapeze and others are in this class.
There are plenty of other factors to look for. What kinds of technology do the vendors use to support wireless roaming? And if they have a wireless switch, do you need to have an actual cable from that switch to each access point, or can you tunnel those links through the corporate network?
These questions will get progressively answered as enterprise Wi-Fi moves from infancy to childhood.
