Andrew Seybold is back.
His commentary in Wireless Week today, Municipal Wi-Fi is creating a Real Mesh, pooh poohs the viability of current “city clouds”.
Wireless Week and other publications have written extensively about muni Wi-Fi. These networks do seem to be the new rage. I wrote a series of commentaries in which I pointed out how unrealistic these systems are when it comes to the number of access points, in-building coverage and other issues.
I do not doubt that the technology needed to build these muni wireless broadband systems exists – mesh network technology, backhaul carried over wired and wireless services (WiMAX?) – so I do not have a problem with the technology. What I do have a problem with is this is happening on unlicensed spectrum over which the city and/or its contractors have no control.
After my commentaries appeared, I received dozens of e-mails from people in San Francisco and Philadelphia telling me they could already “see” 10 or more access points from their homes and/or offices. Several said the range of their own access point had become limited because neighbors on all sides also had deployed access points for their own use.
This is my issue. If you plan to spend millions of dollars on muni Wi-Fi, you will have to spend millions more to keep it running and to fix it when it breaks due to new interference. You cannot simply sweep an area, find places where there is no interference and install access points. There is no guarantee that tomorrow there won’t be new access points that will affect your system.
Seybold, who consults for the cellular industry, makes some good points. Of course Ron Sege, CEO of Tropos Networks, the company who has installed more municipal mesh networks than anyone, disagrees.
All Wi-Fi access points in a given area fairly share the available bandwidth. As more individuals use the access points, they share the same amount of capacity just as they do when using a cable modem and network to access the Internet. While capacity per user is reduced, range is not. In many cases, more access points will be needed over time, not to mitigate interference but to create more capacity as network use grows.
Rather than a technical flaw that undermines a business case, as Andy suggests, such scalability of nodes and gateways is actually a feature of an efficient wireless mesh design. Unlike DSL, cable or cellular networks, mesh networks are flexible and can be built out in stages and redesigned to respond to growing subscriber demand, rather than an all or nothing.
“I don’t believe it is possible to ‘blanket’ a place like San Francisco with Wireless LAN, and certainly not with the 300 access points per square mile Google is suggesting”, opined Arraycomm’s Martin Cooper, during a session he chaired at the Broadband Worldwide forum in Madrid recently. “I calculate they will need more like 3000 AP’s per square mile.”
Google has suggested it can provide 95 percent coverage outdoors and 90 percent indoors with the 300 Wi-Fi Access points.
“They are clearly a great company, but I suggest they need a few lessons in wireless propagation,” said Cooper.
Of course, Arraycomm has their own solution; beamed WiMax. I think he also meant to say 30 access points per square mile [a generally accepted figure] not 300.
Whatever.
.gif)
