Q&A with Kenneth Carnesi on the Harvard Square WiFi hotzone, what went wrong with muni WiFi and more
Last week, I spoke to Kenneth Carnesi, CEO of Anaptyx, a Boston-based firm that deployed the Harvard Square WiFi hotzone. Anaptyx continues to run the hotzone as its service provider. Since they began offering free WiFi service in Harvard Square in June 2008, 83,068 people have used the network.
Here is my Q&A with Kenneth. I thought it would be interesting to know from the service provider’s perspective what surprised them about the network, what common mistakes are made by those who are new to outdoor WiFi deployments, whether the developments in 802.11n networks will make a huge difference in outdoor WiFi networks, what went wrong with many municipal wireless projects and what the federal stimulus funds can mean for broadband in the United States.
(1) What are the most surprising things you’ve seen from the Harvard Square WiFi deployment which you did not expect?
Apart from the tremendous number of iPhone/iPod users we see on this network or most other municipal networks (close to 50%), I continue to be surprised by how fast the network has grown. On average, every seven weeks 10,000 new devices access the network. I see tremendously fast adoption rates in nearly all of our networks, but the density of users in Harvard Square really just makes for some huge numbers.
(2) What are the mistakes that people, who are new to deploying a large WiFi hotzone, make?
The most common mistakes I see individuals make when deploying a large WiFi network are related to planning and hesitation. On the planning side it’s a multi-faceted issue, boiling down to three main concerns: budgeting, performance and overall expectations. These three topics ultimately tie into each other. When deciding to build a large network it is necessary to first set the expectations of this network: Are you going to support VOIP? What areas of the network will you see the highest concentration of users? What speeds are you looking to provide? All of these questions and more will determine exactly how many access points you will require and where your gateways (main Internet injection points) should be placed.
After determining the answers to these questions, either yourself or an engineering/deployment team like Anaptyx can determine what your budget is going to have to be. From that initial figure you can then either raise or lower your expectations to meet your budget or vice-versa. Keep in mind that part of the budget needs to be the ongoing expenses to operate the network, such as data line and support contract costs if you are not planning to support the network yourself.
As far as the second concern, hesitation, many people are simply hesitant to put in a medium- to large-scale (municipal) network. I think much of this stems from the costs and failures seen from the majority of what I consider “First Generation” networks. In the last two years alone technology has evolved so quickly especially with mesh technologies like Meraki that the cost to build a successful network is very affordable even for the most budget conscious owner. Trust me, if you build it they will come and you won’t break the bank doing it.
(3) What advice would you give to municipalities and other service providers based on your experience at Anaptyx?
I don’t want to up sell here too much but I would certainly advise them to work with an engineering/deployment firm when it comes time to plan, design and price the network. All too many times I have seen individuals who attempted to build the network themselves without any design or deployment expertise and often with standard, consumer grade wireless equipment available at your local Best Buy. Don’t ever do this; hire a professional in some capacity. While the initial cost may be higher, it is much cheaper to start properly from scratch than it is to build a network twice.
Other tips, design wise, is to ensure the network is designed in a modular fashion so you may add-on or make changes as necessary down the line. Also, once you have identified the areas with highest user density (these are typically parks, squares, centers or other areas of “congregation”) focus on these areas by adding multiple access points. Sometimes just because you have good signal doesn’t mean the network has been properly designed to take into account the fact that there may be too many users logging on at once in a single area for one access point to provide enough bandwidth.
(4) How do you get a lot of bandwidth into the Harvard Square network to satisfy the users’ needs? I am assuming that most of the users are students who use bandwidth-hungry applications, e.g. streaming video and more. Do you have a fiber network as backhaul?
In Harvard Square we are providing bandwidth through a mix of commercial grade coaxial and copper (DSL) data lines. We have 32 megabits total by way of coaxial and an additional 14 megabits of download speed available by way of copper. As far as usage goes, many students and younger users in general are using the network and out of the roughly 7,000 users we saw last month nearly half of them were Apple iPhones or iPod Touch’s (versus about 70 Blackberry’s/Smart phones). These devices rarely download more than 20 megabytes or so per month because they are mainly used for simple web browsing and e-mail. As for the rest, about 1,000 were Windows users and 1,700 were Mac users – most of these users transfer between 100 and 700 megabytes per month, which would be typical web browsing, e-mail with attachments and few YouTube and probably iTunes downloads here and there. Out of all our users, only 1% or 2% are transferring more than two gigabytes per month. Interestingly, 3% or 4% of our devices are gaming systems (PSP, XBOX, PS3, Nintendo) or Slingboxes. We don’t actively block these users, but we do have a splash page reappear every so often to curb uninterrupted connections for an extended time.
(5) Why did Harvard Square decide to use the solar powered Meraki units? Was it only to save on energy costs? Or it is because solar units allow you to place the wireless access points anywhere you want?
The decision to go with solar powered Meraki units was made mainly because of cost. There were a few locations in Harvard Square where wireless access was essential, yet the cost to bring an electrical line up to or through the roof of a building (where we mount the APs) was very high. In these cases, it was cheaper and easier. It’s also very cool.
(6) Who is paying you for the network deployment and maintenance? Is it the Harvard Square Business Association?
The Harvard Square Business Association is paying for the network deployment and maintenance. The association is really an ideal client. Denise Jillson, the Executive Director of the HSBA, had a great vision a few years back when we initially deployed the network and I am glad she found us. The network has been a huge success and the HSBA’s ability to be in tune with technology and aware of the fact that the network does have ongoing maintenance and expansion costs makes them great to work with. I think they were the first in the area to take a leap like this and it really paid off as far as the tremendous value they get from this successful network in both advertising and driving visitors to Harvard Square.
(7) What other significant Wi-Fi projects has Anaptyx undertaken? What projects are currently under deployment?
We operate in the commercial, municipal and residential industries, but as far as public outdoor deployments go, Anaptyx has completed networks in Harvard Square, Nashua (New Hampshire), Harrisville (New Hampshire), Readers Park in downtown Boston and Dorchester (Massachusetts). Currently underway we have projects in La Plata (Maryland), Newmarket (New Hampshire), Kutztown Borough (Pennsylvania), and Middlebury (Vermont) and are working on a public safety pilot spanning 1.3 square miles for the City of Boston.
(8) On the technical side, are you currently using any of the latest 802.11n outdoor mesh access points (whether from Meraki or other vendors)? Have you seen dramatic improvements in speed and latency versus the old 802.11b/g networks? What is the most compelling reason for a service provider to move to 802.11n?
We certainly have used 802.11n outdoor access points. Although we do work with other vendors from time to time, Meraki is very often the ideal solution for the types of networks we deploy. In any case, 802.11n is definitely an improvement over b/g as far as latency, performance over hops and overall user capacity are concerned. However, the 5 GHz band and multi-radio setups are something I think many service providers overlook. In the simplest terms, 5 GHz radios — when used for backhaul purposes — enable us to get over much of the interference in the 2.4 GHz band (many times the interference can even come from your own network). Proper use of 5 GHz will ensure higher throughput and stability for backhaul, point-to-point and point-to-multipoint purposes.
Separately, a multi-radio design enables you to separate backhaul/mesh (traffic between access points) from client traffic (access points to laptops). This takes a great load off the access points, enabling them to serve more clients simultaneously and it tremendously increases throughput over multiple hops. The unique feature of Meraki’s MR58 that keeps us going back is its ability to combine 802.11n/5 Gigahertz and multi-radio design into a single package while still retaining the capability to operate as a mesh/cloud server access point. In my opinion, this makes it the ultimate radio for large-scale outdoor use.
(9) In your opinion, what went wrong with municipal wireless projects?
(Sigh) I don’t even know how to get started here. I think that, on the most basic level, these networks were built at the wrong time to be building municipal networks and with the wrong intentions in mind. I won’t name any names, but to put it into perspective, some of these networks were built when ZIP drives, Monochrome Cell Phones and 30 Gigabyte hard drives were all the fad. There was no way these networks would see the kind of adoption they had been planned for – hardly any devices had WiFi chipsets in them. Coupled with poor business plans, blanket coverage, a lack of advanced mesh technology, and the fact that every access point was a multi-thousand dollar expense, it was a recipe for disaster.
Today, almost every consumer device imaginable has an embedded WiFi chipset; second and third generation mesh technology is readily available at an affordable prices. This enables us to more easily deploy focused or concentrated hotspot style networks wherever they are needed, rather than providing blanket coverage. Often the investment is minimal enough that providers are able to supply free wireless Internet access of some capacity, and they can many times turn a profit or at least cover operating costs through advertising and community involvement.
(10) Do you think that the federal government’s broadband stimulus plan will make a significant difference in bringing broadband to underserved areas? What is your opinion of the broadband stimulus plan? Do you think there should be a broadband stimulus plan for urban areas?
I certainly hope so, because it is needed. In my opinion, however, the broadband stimulus plan has made it extremely difficult for applicants to adhere to their strict policies on qualifying deployment areas, especially for those with little or no budget. The main issue stems from the providers and a lack of information. Giant providers like Comcast or Verizon haven’t (not surprisingly) provided anyone with their user demographics. They can easily close out an area as “unserved” so long as they can meet the ridiculous government definition of “broadband” or provide an affordable data plan that meets this requirement. As far as information is concerned, applicants are asked to break down applications by census block, yet no information is provided by the government or any other entity (and certainly not for free) that breaks down adoption rates, service providers, etc. by census block.
The government needs to arm applicants with more information or, in my opinion, provide qualified applicants with a small yet appropriate budget to work with pre-screened professional engineers or deployment firms who are capable of providing a realistic network design/layout and appropriately respond to the application. Unfortunately, without this help, many applicants simply cannot afford to respond properly even though their project may be perfectly qualified for funding.
As far as funding for urban areas, I believe this is absolutely essential. Although broadband may sometimes be readily available in most urban areas, these are also typically the areas with the highest concentration of children/students, families and individuals in general who have higher costs of living and a great requirement for Internet access — yet they have been affected just as much by the economic downturn as anyone else in this country and are now less able to afford broadband access. These are facts not evident through old census data. Our economic crisis is fairly recent and a true, nationwide, targeted census of broadband access has yet to be completed. Urban areas, in my opinion, have the best opportunity of any deployment area to take a last-mile (BTOP) funded deployment and make great use of it by tying the network in with sustaining broadband adoption grants, community health centers and computer learning centers. A great example of the potential to tie WiFi access in with the community is the work we have completed for Dotwell, a nonprofit in the Dorchester area of Boston and the Cambridge Housing Authority.
About Kenneth Carnesi: he is the CEO of Anaptyx, a Boston-based firm that has deployed outdoor WiFi networks in New England.
* * * * * * * * * * * *
Buy these Research Reports now:
(1) Guide to the WiMAX Band (2.5 GHz): the technology, license holders and future prospects
(2) The U.S. Mobile Web Market: Taking Advantage of the iPhone Phenomenon
* * * * * * * * * * * *
Related posts:
Loading ...
