Santa Clara, CA muni Wi-Fi network gets a second chance

I know you’re thinking what I’m thinking: “Here we go again.” Santa Clara, California has announced that it is re-launching its free citywide Wi-Fi network, only this time, it’s not the network created by MetroFi (the bankrupt wireless ISP that had contracts to deploy Wi-Fi networks in various cities across the US) but the city’s own.

Background: MetroFi had won a contract to unwire Santa Clara, but the company went into bankruptcy after it failed to make money. Santa Clara bought the network (the MetroFi access points) from MetroFi and issued a public tender to find a systems integrator to build a wireless automated meter reading system. Silicon Valley Power (SVP), the city-owned utility, chose Elster Solutions to build the AMR system, add more Wi-Fi access points and at the same time, create a citywide Wi-Fi network for use by anyone in Santa Clara.

The city says the network should work better than the old MetroFi network because it has a greater density of access points (40 APs per square mile versus MetroFi’s 30 APs) and it delivers 2 Mbps versus MetroFi’s 1 Mbps. Moreover, the city says users can get the Wi-Fi signal indoors as well.

What’s the obsession with getting the signal indoors?

Having heard this before, I’m skeptical because I’ve seen how difficult it is to deliver Wi-Fi indoors. I do not think goosing up a network to deliver indoor Wi-Fi everywhere is necessary at all because people have Wi-Fi at home and at work. What matters most is that the outdoor service is fast, reliable and easy to log into since most people using it will be on iPhones and iPads. Since Santa Clara is largely suburban, it makes more sense to focus Wi-Fi access in those places where people congregate, e.g., parks, public squares and shopping malls. Is the indoor free Wi-Fi service available inside shopping malls? If yes, that’s great. If not, why not?

Is 40 APs per square mile enough?

I have doubts about whether 40 access points per square mile are enough. Ken Biba, who has deployed many wireless networks, advises 60 access points per square mile (see Guidelines for Creating Outdoor Wi-Fi Networks). If those 40 access points includes the old MetroFi access points (instead of completely new ones), it is highly unlikely that the network will be able to deliver the speed and coverage promised by the city, and it’s even less likely that the people sitting inside homes and commercial buildings will be able to access the network. Old 802.11g based Wi-Fi networks deliver terrible service to most laptops and smartphones. I hope SVP replaced the old MetroFi APs.

There are two things this network has going for it:

(1) The network owner and operator, Silicon Valley Power, is using it for wireless AMR as well, which means the network has a dual function — public Wi-Fi access and automated meter reading.

(2) SVP has an extensive fiber optic network that it uses as backhaul for the Wi-Fi network. Many Wi-Fi operators have to pay for the high costs of backhaul; SVP does not.

I hope SVP makes this network a model for all other cities and counties that are still exploring ways to use Wi-Fi not only for municipal applications (such as AMR), but also for use by the public whether directly (as SVP has done) or by wholesaling capacity to ISPs and mobile operators (which is what Towerstream is doing in NYC).

Required reading:

Guidelines for Creating Outdoor Wi-Fi Networks

Related articles:

Santa Clara uses old MetroFi network for automated meter reading

Santa Clara public utility issues RFP for wireless AMR

 

Comments

  1. doug berman says:

    Having deployed hundreds of outdoor AP networks. I can say that 40 AP’s per square mile was challenging. It is questionable whether deploying 60 AP’s per square mile will make the user experience better. This is because deploying high power APs in such close proximity to one another will cause self interference and reduce available bandwidth, regardless of capacity. Even with sophisticated power techniques, you will never be able to predict where clients are located, their signal strength and their WiFi driver characteristics. I agree with your statement; getting WiFi indoors should not be the goal; but rather the exception. Deploying WiFi in strategic areas such as malls, parks, marinas, and downtown areas should be the focus. It would be wise to focus on meter reading first; than, focus on Internet access in strategic areas.

  2. We are really excited about offering free outdoor Wi-Fi as a by-product of our advanced metering, smart grid and smart city effort. I fully appreciate the commentary and want to correct the misperception that we are attempting to provide indoor service. We are not and have never intended to do so. With a customer premise repeater – such as those made by Ruckus – many customers have been able to move the signal into their homes and businesses. That fact was misconstrued into the idea that we were providing indoor service. Providing connectivity outdoors is the goal of turning up this function in our utility-grade wireless WAN.

    Secondly, 40 APs per square mile is the build design to provide a utility-grade, 802.11B/G/N WAN. This will support several applications for mobile workforce, support public service, provide WAN backhual for distributed monitoring and control and more. Whether or not this density will provide a “peak” experience for consumers is not well known and there are many, many variables. Also, 40 APs per square mile is a design starting point for our residential areas with a lesser density in our commercial industrial areas. We are already anticipating that some amount of infill will be necessary once the system is up and tuned to meet our own needs and respond to clear opportunities within our consumer base.

    A second misperception I have seen out there is that consumers will have a roaming capability where they will not be dropped while (passengers) are using their Wi-Fi connection in a moving vehicle. We will be testing mobile 802.11N routers in fire department vehicles for this capability, but this is not a design criteria of the system for normal users.

    Thanks for keeping an eye on this. We benefit greatly from the discussion.