802.11N promises to lift the muni Wi-Fi market

Novarum’s latest ratings on municipal Wi-Fi networks are out and, not only do they find that muni Wi-Fi is “far from dead,” they found that muni Wi-Fi networks can operate as reliably as cellular data services if next-gen 802.11N client devices or high-performance adapters and high-gain antennas are used.Novarum’s latest ratings on municipal Wi-Fi networks are out and, not only do they find that muni Wi-Fi is “far from dead,” they found that muni Wi-Fi networks can operate as reliably as cellular data services if next-gen 802.11n client devices or high-performance adapters and high-gain antennas are used.

These findings come as great news to an industry that’s been questioning its business models since August when EarthLink announced its aproach to the muni market was unworkable.

Since that time munis have been bludgeoned by press reports that question everything from the business model to the viability of the technology itself. Novarum’s report takes on the last half of that question and shoots down the criticism that current muni Wi-Fi networks will be rendered obsolete with the introduction of 802.11n devices.

Novarum, which ranks metro Wi-Fi networks according to performance, availability, ease of use, and value found that 802.11n client devices not only worked on current muni Wi-Fi networks, they were able to pull signals where traditional 802.11 a,b & g devices were not and worked without difficulty in areas where traditional devices registered weak signals.

The company also compared the performance of the a, b, g, and pre-n devices to devices equipped with high-performance adapters and high gain antennas and to 3G cellular services in the 16 metro markets it tested.

In a press release accompanying the announcement of the data this morning, Novarum co-founder Ken Biba, declared muni networks “far from dead” and noted that, although metro Wi-Fi networks generally end up costing more than what the political hype surround them suggests, they “may indeed deliver a higher quality of service at a lower cost than other wireless broadband alternatives.”

Novarum is making the summary report available, as well as city reports on Metro Wi-Fi and cellular data services in Philadelphia PA, Portland OR, Rochelle IL, Santa Clara CA, Tempe AZ and Mountain View CA.

Novarum co-founder Phil Belanger told me that they found that 802.11n devices demonstrated good range and good performance. In areas where standard 802.11b and g products failed to pull signals or the signal was weak, the N devices typically often had no trouble. “In this application we’re not going any faster,” he said, “but rather we’re talking about signal availability. We found this in Mountainview. We also found this on every other network we tested. It shocked us, really, how well it worked. We think that’s huge for the industry. A network that appeared to be marginal seemed solid when you used an N client.” Mobile applications using high-performance adapters and high-gain antennas achieved even better results.

As competition to cellular data services, the report found Metro Wi-Fi delivers twice the performance. According to Belanger, the difficulty with muni Wi-Fi deployments is not the speed; it’s the availability of the signal. “Forty nodes per square mile are not enough,” he said.

Other key findings in the report include:

  • Cellular data services improve performance – the deployment of 1xEVDO Rev A and HSPA increase performance by 30% overall and more for uploads.
  • 3G cellular data services are still not ubiquitous in the North America. True 3G had a 60 percent Service Availability in the cities tested.
  • 3G cellular data services use 2G service as a backup. Combined Service Availability is 87 percent in the cities tested, making cellular data services the availability leader.
  • Novarum announced its “best of” wireless awards for 2007. Toronto’s OneZone once again weighed in with the best overall Wi-Fi service while Verizon achieved “Best Overall Cellular Data Service” in Mountain View, Ca. EarthLink’s Feather network in Philadelphia got the award for most improved Wi-Fi network; since Novarum’s last survey, Philadelphia increased its node count from 32 to 48 with an accompanying improvement in availability from 70 percent to 85-90 percent.

    Click here to access Novarum’s reports.


    1. Lots of great analysis here from Carol Ellison and Novarum.

      I agree that 802.11n holds a lot of promise for the municipal broadband market. Cisco, Tropos and other big infrastructure providers have endorsed the current 802.11n draft standard, which should give customers peace of mind. Cisco announced its 802.11n push in September. Tropos followed with a D-Link relationship in early October.

      But the municipal broadband industry has to be careful. Too often we believe a specific technology upgrade can solve a lingering business problem. Yes, 802.11n will solve some coverage, connectivity and throughput issues. And standards from the IEEE have helped to accelerate IT markets. But we as an industry have to continue to identify successful business models for municipalities to pursue.

      Without clear broadband business models in place, some municipalities won’t be able to find the dollars needed for 802.11n infrastructure.

    2. Joe,
      You are absolutely right. The last thing the industry needs is a new slogan that sets false hopes for a rosier future. Novarum is certainly not saying that 802.11n will solve the problems of the industry. We are simply documenting a bit of positive news from our testing reports. The early tests indicate that 802.11n clients work better on the existing Metro Wi-Fi infrastructure. It is nice to be able to report some positive news in this industry.
      (There is negative news in the reports too. For example, most of the current Metro Wi-Fi networks are deployed too thinly to deliver quality service to their targeted clients.)

      So what is the real impact of better performing 802.11n clients?

      For consumers it is another reason to upgrade to 802.11n. 11n is compatible with current networks; and it works better on current networks in the home, in the office and now all over town.

      For Metro Wi-Fi equipment vendors, it is exciting to design this technology into their new products. It appears to work better than we were expecting in this environment. Imagine new PepLink, Ruckus or Meraki boxes based on 802.11n. How will Tropos, BelAir, Firetide, Cisco and Strix leverage this technology?

      For cities and service providers, 802.11n is not a free pass by any means. Any Metro Wi-Fi network that supports public internet access is also an open network. This means ANY Wi-Fi client could be using the network. The big challenge for open Metro Wi-Fi networks is that they must support a wide variety of client devices with very different performance capabilities. There is no performance certification in the 802.11 industry. The Wi-Fi Alliance does a great job of interoperability certification, but the performance bar for Wi-Fi certification is very low. So an open Metro WI-Fi network has to support the worst case Wi-Fi client devices, and their radio performance may be very poor.
      (This could be the topic of a very long standalone article… One advantage that the closed cellular networks in the US have is that they can and do control the performance capabilities of client devices using their networks.)

      So 802.11n is positive news for cities and service providers. A new generation of client devices will have a better experience on their networks and a more positive perception of these networks. However, they still have to figure out how they are going to support the less capable Wi-Fi clients. If they want to provide the same coverage and service level guarantees for all Wi-Fi clients, that will require a very dense infrastructure – lots of APs per square mile. And that takes us right back to the business model discussion.

      A big variable in the business model discussion is the cost of the infrastructure. You need to understand the applications and client devices that will be using the network BEFORE you can figure out how much the infrastructure will cost. 802.11n is just one ingredient in the recipe for a successful wireless broadband business model.

    3. In the Unwire Portland RFP, included by reference in the Nonexclusive License Agreement that MetroFi signed, section says:

      “Any individual using a device containing an integrated/built-in IEEE 802.11b/g (MiniPCI) radio or non-integrated/external card (PCMCIA, USB, etc.) shall be able to receive acceptable signal and mostly uninterrupted service in all outdoor Coverage Areas.”

      and section says the same thing. It seems to me the only reasonable interpretation of those clauses is that MetroFi agreed to provide connections to anyone using 30 mW b/g client gear with low-gain antennas, which were widely extant at the time of the RFP and Agreement and remain extant today. I have yet to see MetroFi acknowledge or address this issue.

      It is important to remember that no amount of speculation or even demonstration that 802.11n client gear does better than the b/g client gear will make the b/g gear do better.

    4. Good comments but just have to throw in my two cents.

      I live in Santa Clara (Metro-Fi) and am 300 feet from the AP. The signal quality is so poor that I can rarely follow any hyperlink without having it timeout. Even if I take my laptop out to the street with direct LOS to the AP, the signal is still poor.

      So, to really use this service when I am roaming, I need to buy a new 11N client adapter and/or high gain antenna and carry all this stuff with me to get a signal at the mall or some other location?

      I pay $18 for ATT DSL @ 3 Mbps and 8 POP3 email addresses. Given the useless performance of Metro-Fi, I am happy to wait until I get to work or home to have Internet access.

    5. It is absolutely the case that one should expect using current technology 802.11g adapters to get access to these networks.

      Novarum tests using exactly this client (as well as others to test both the technology and other applications – such as public safety) and so all our tests are at core using the SAME client – allowing us to compare apples to apples.

      The bad news is that many of these early WiFi networks were designed with poor business models that did not allow for the necessary investment in infrastructure necessary to give a good competent service with these client devices.

      Portland and Santa Clara networks are deployed at roughly 30 nodes per square mile and frankly do not give very good coverage with these 30 mW miniPCI clients.

      The good news is that some networks do give good coverage AND performance. In our testing we particularly note Toronto ON and St. Cloud FL as networks that do work well.

      And the other message is that it is also clear that networks can get better – with investment. Philadelphia was initially a disappointing network. But in the six months between our first test and our second … Earthlink increased the node density by over 50% and that had a dramatic improvement on the availability and useability of the service.

      It is frankly still early days for this technology – and one of the key messages is that all wireless technologies will have teething problems – and that we have already seen that is possible to build a good network and that we are beginning to gain insight into HOW to reliably design and deploy these networks.

      Expect similar teething challenges for mobile WiMax …

      And expect wire problems. I just returned home from a two week trip (testing new networks) to discover BOTH my ATT DSL AND Comcast cable connections down with wiring problems outside my home. Comcast responded in one day … but ATT still is not working after 4 days. YMMV.