Slashdot Mirror
How Much is Riding on Wi-Fi?
nexex writes "The Seattle Post-Intelligencer's John Cook explores the current flood of money on wireless networking startups and if they could be heading towards another dotcom bubble. Interesting tidbits include, ;More than 60 Wi-Fi start-ups have raised more than $650 million in the past two years, according to VentureWire. Last quarter, there was more money invested into wireless technologies than networking and enterprise software.'" The article's got some good commentary on grassroots-founded tech trends vs. investment-backed tech trends, and tries to explain why wi-fi has caught on so well.
Comparing this to the dot com bubble is not really a very good comparison. No one really needed thousands of websites which attempted to sell you services that you could get from any high street shop. But I can see genuine uses for wireless technology which means that it is worth investing in.
Cause I just noticed that local used office equiment store is running low on inventory.
This space for rent.
I've dealt with several wireless startups in my career. I feel the analogy of the dot-com bubble is quite accurate. Let me explain why...
One of the more prominent start-ups involved in building gear was requesting an ASIC (application-specific IC) engagement from my company to implement some of their functionality. They were asking about how fast we could run a certain type of embedded processor in one of our processes technologies. Keep in mind that these are guys with supposedly years of ASIC and system-building experience. When I turned the question around to them that it was more relevant for us to provide a solution with a certain amount of floating point performance, MIPS, multiply-accumulates per second, I/D cache size, etc., they kept saying that they weren't sure, but that clock speed was paramount. Yeah, right...
To top it off, they wouldn't give us any details of their end application. Was it 802.11a/b/g? 802.16? On-board multipath antenna signal processing? They also said if we asked too many questions we'd be out of the running for an ASIC bid. In other words, there was little substance to what they were dealing with. Yet, they were supposedly one of the most promising companies out there.
Then I took a stroll through the Bluetooh forum a few months ago in San Jose, CA. I saw a lot of folks involved in wireless IP not just for Bluetooth but for 802.11. Based on this, and my experiences with companies as described above, my verdict on wireless is as follows:
1. There are too many players who don't know what they are really doing, and who have no focused strategy. They're just getting into wireless because it is the industry's newest buzzword. That's at all levels of the value chain (semiconductors, box builders, and service providers).
2. There are far far too many players in the semiconductor aspect of wireless. From soft/hard IP providers to chipsets, it's a confusing soup whose interoperability is unconfirmed, and who are jockeying for position on issues such as range, power consumption, and how integrated they are (both from the baseband+PHY perspective and from the driver/software stack perspective). In some cases, the IP hasn't been tested or even implemented in an FPGA, yet they're on the show floor peddling their wares. There'll be a major shake-up in this area not only because of oversaturation of players, but because of oversaturation of silicon suppliers, where profit margins of the manufacturers are being pushed almost endlessly downwards due to overcapacity in semiconductor manufacturing and desperation of some companies to stay in business. Most of these players should disappear and leave us with hopefully two or three good standard chipsets per major standard group. Those looking at integreated wireless ASICs with PHY are only dreaming for the next several years.
3. In the system arena (commercial/residential wireless APs, repeaters), everyone is jumping on the bandwagon. Yet, as shown by the company I described above, there is a headlong march to get these products out without looking at some of the fundamentals such as interoperability. Heck, I had a friend yesterday whose Linksys PC card wouldn't link to her Netgear AP. That's a tiny example, but we could potentially be facing some of this type of problem.
4. In the service provider arena, there are some revenue opportunities. The end market, however, needs to have greater uptake of compatible wireless gear. That's going to be very difficult. There's only a limited amount of bandwidth available in the already-crowded space. For example, 2.4GHz is for 802.11b/g, and that's already crowded with devices from cordless phones to microwave ovens that could be potential sources of interference. If wireless is to be successful commercially, as a service, I think we'll either have to piggy-back on the 3G networks, or set up a standard that doesn't use frequencies fully opened up by the FCC. Of course, you know what that could mean (the big fis