Archive for May, 2008

On May 7th, this article appeared in the Philly Metro saying,

“Earthlink, which stopped accepting new customers last week, has given the city until tomorrow to come up with a plan to take over the system or it could begin to take down the network, according to sources close to discussions. An original deadline of last Wednesday came and passed.”

On May 13th, this article states,

“The company said it is providing customers in the area with service through a 30-day transition period that ends June 12. It is getting in touch with these customers to give them information about the termination and help move them to other EarthLink Internet services.”

In the interest of making sure that analyzing what took us from a brilliantly conceived concept to the shambles we see today doesn’t turn out to be the most valuable asset of this entire affair I would like to present the following perspective.

What happened when –

At the onset, no one could have predicted that inquiring how much it would cost to build a wireless network covering all of Philadelphia would spark a nationwide debate and give birth to a new industry. What went from a proposal to build the network for $10 million, then a $5 million counter proposal, to Earthlink proposing to build the network at no cost to Philadelphia, has provided an exciting ride.

When the details of Eathlink’s plan was first announced, I went on record saying that the network wouldn’t work, it couldn’t work. This realization had nothing to do with me having any particular prophetic powers, perish the thought, the basic numbers just didn’t add up. In fact, what few numbers were being released not only didn’t add up, they appeared to show that no one was paying attention to the business model. While that aspect was disturbing, when it came to the network engineering, the originally announced design caused waves of outright derision from the Wireless Internet Service Provider community – the very people Wireless Philadelphia and Earthlink should have been engaging in this discussion.

For those of us who had actually built these networks what was publicly being stated was simply not possible to deliver. Of course, this may seem easy to say from today’s vantage point but the record is out there preserved in the internet archives for anyone who cares enough to read through it all.

Certainly, many things went right and this, in of itself, is astounding considering that there was no blueprint to work from, no established baseline or best practices to adopt. The idea that Digital Inclusion would need to be an integral part of this network was something that hit the mainstream, as did the incredible job of holding public meetings in an attempt to educate the entire city as to what they might expect. The work that Dianah Neff did was nothing short of Herculean and regardless of the outcome, moving forward, she should be lauded for managing to accomplish what she did. There were also countless others, including Karen Archer-Perry, to name only one, who worked miracles to move this project along and there needs to be recognition for their work, as well.

At the same time, Wireless Philadelphia’s reliance on Earthlink “taking care of everything” while being incredibly seductive, was naive, to put it nicely. And to be completely blunt, if the mantle of failure has to be accepted by anyone, Earthlink needs to step up and face the facts. Yes, the network design was handed to Motorola (who chose to partner with Tropos) and while we should be looking at these two companies as having some responsibility in this mess, Earthlink’s name was on the project and they should have made sure that their project would do what they were promising the world. There was also a pilot project built using Alvarion equipment that performed very well but was not chosen, why, I’m not sure we’ll never know.

We’re here. What now? –

Certainly, everyone has an opinion. Ethos, published a report, which can be read here, describing their vision and how they would address the problem, a solution I do not hold as being desirable or sustainable.

While I will not attempt to put forth a 64 page report here, or even try to explain why I don’t believe Ethos is on the right track, I will state, very simply, that there is a workable plan, based on technology available today, that could take this misstep, learn from the mistakes and build out a world class network from there.

First off, let’s clearly articulate four mandates that must be adhered to if this, or any such project, is to succeed.

    1.) The network must be technologically sustainable.
    2.) The business model must make financial sense.
    3.) There must be compelling reasons for the customers to buy in.
    4.) The plan must include a long term process for continual improvement.

Technological Sustainability –

There is an accepted wisdom in this industry that bandwidth demands will continue to increase, at an ever-increasing rate, for the foreseeable future. Based on that wisdom, any network built, regardless of technology employed, will be temporary and as newer products are introduced they will need to be integrated into these networks. In the license exempt wireless industry, the useful lifespan of any equipment is generally considered to be three years. With respect to the fiber industry, this period of usefulness is extended to a decade or more, which should be expected based on the cost differential between these two technologies.

One fact needs to be clearly understood, in order for any network to continue to be indispensable, there will always be two distinct components; “fixed” and “mobile” connectivity.

Business Model –

Borrowing from a century’s worth of experience that the telecommunications industry has accrued, one well documented reality must be adhered to – businesses will pay more than residential customers for a business class service. This means that if you want to build a business model that will work, you need to have services that the business community will seek out and purchase. This also means is that “best effort” WiFi technology is not going to meet those requirements, nor is the promise of the same services offered by everybody else in the industry going to differentiate your offerings from the pack.

In today’s competitive world, the bar is set pretty high and any new competitor in this arena MUST offer not only competitively priced services but services that cannot be purchased from anyone else.

Specifically, any such venture must provide a speed/price component that noticeably beats the competition while meeting or exceeding the reliability metrics that anyone else can offer. At this time, the bar has been set so that any entity looking to penetrate this market must be able to deliver a service that equals Verizon’s Fios, which currently delivers 30Mbps symmetrical to those who wish to pay for it. That’s right, if you think you’re going to make a huge hit in any market by offering 5Mbps, (forget 1.5Mbps or dialup 2.0, as I like to think of it) you need to find a new business. However, this is the entry point for fixed communications, to succeed you will need to also be able to deliver a five 9s reliable, competitively priced, connection that can reach multi-gigabit speeds and this service must be installable with a very short window of time – as in three business days.

Additionally, make this service portable, meaning that wherever the customer goes, the service is available, albeit with the understanding that multi-gigabit speeds will not follow them everywhere. So, when customer “A” leaves their office, their wireless device will continue to keep them connected to the network. When this same customer heads home (assuming that they lives inside the network’s 100% citywide cloud) their service follows them there.

There is another component to the revenue stream that must be included in order for this business to thrive, applications. While an entire book could be written about this subject alone, it is the applications that are provided on these networks that create the demand – and if you aren’t supplying more and better applications than the competition, why would your potential customers seek out your services?

Compelling motivations for adoption –

As mentioned above, applications that are only available on your network will oftentimes sway the customer in your direction and may be an even more persuasive argument than price/performance alone.

Without getting too far into the subject, applications comes in two distinct categories; revenue producing and quality of life enhancing. Three revenue producing applications that are commonly paid for currently are voice (telephone/cell phone), Internet service and television/video entertainment. It is the author’s belief that in the next 24 to 36 months voice and video will cease to produce revenue, (in any meaningful manner) and will become part of the value added when one subscribes to Internet service. Please adjust your revenue projections accordingly. In place of those applications, I see video security monitoring, remote medical diagnostics and monitoring as well as mobile finance coupled with location based advertising as being some of the revenue producing applications which will supplant voice and video revenue streams.

Turning to the quality of life applications, the ability for a parent who is unable to attend their child’s activities, such as a baseball game, recital or school play but who can remotely attend these events is a compelling argument for adoption of a service. This same rationale holds true for a parent who would like to be able to check in on their child at day care or check on an elderly parent at home.

To highlight a valuable business application that could be implemented (and is not available on the competition’s network) I heard a presentation given by a VoIP provider who provides a service that allows small businesses, ones that might only have two or three lines, to “stack calls” during times when their line capacity is exceeded. As an example, he mentioned that in a college town, he had a customer that ran a pizza delivery service. Under normal business loads two phone lines was all that the customer required but when there was a sporting event or on weekend nights, just before closing, the call volume went through the roof. To answer this need, this provider created a service where when a customer called, and all the lines were in use, instead of receiving a busy signal or being shunted off to an answering service, the call was kept alive with “on hold” music until a live person could answer. This allowed the business to handle more volume than they could otherwise and minimized the occurrence of callers that would have normally dialed the competitor. The cost for this service? A mere quarter per call! While this was a service that was certainly a bargain to the business owner, it also provided a healthy revenue stream to the service provider, especially when all of the competing businesses in the area subscribed to this service.

One more application that is network agnostic but should be adopted into the offerings presented by any savvy network operator, Telepresence. While you may not be familiar with this term, eventually this will be a must have service for any business that used to travel in the past. The ability to hold a face to face meeting with someone without leaving your office (or perhaps, only having to travel to your friendly, local Telepresence center) will become more commonplace as the cost of travel rises – and I don’t know anyone that is predicting the costs will significantly decrease in the longer term.


Continued improvement –

And now we get to the least palatable aspect of this discussion.

Simply building a network and marketing it’s services is not going to cut it, in case you might have thought otherwise.

This network, or any network, for that matter, will need a plan that will balance revenue coming in with the forecast expenditures which will be ongoing forever.

While a detailed plan could also take up more space than I am willing to dedicate here, the reality is this will be a phased project with the next decade planned out.

Phase one –

The network will be built using a fiber backbone that delivers connectivity to all areas of the city with more than one fiber connection provided to each area for redundancy’s sake. This fiber construction has already begun in Philadelphia, and will not stop, remaining ongoing until every single building in the city is eventually connected, perhaps over the next decade.

From the initial fiber buildout, rings of multi-gigabit wireless (and/or Free Space Optics) devices will be deployed to become the next layer of the network. This series of rings will be engineered to allow companies/organizations that want ultra high speed service to be connected FIRSTas they are the premium customers.

Further down, this is the stage where the next layer of wireless rings will deliver connectivity measuring in the hundreds of megabits. From this layer we will then connect the next group of customers that are looking for 30 to 100Mbps service – as they are also a premium class of customers.

At this point, we have reached the level where the “edge” network can now connect to the main network backbone. Please note – this is what Earthlink/Motorola did on a far slower threshold. It should be noted that nowhere in this design is Mesh Technology specified and that the equipment I would spec out to connect at the edge would be very different from what exists currently.

As things stand, this phase of the project would be projected to take eighteen months to two years, assuming that all political roadblocks were removed and adequate resources were dedicated to the task.

Phase two –

Complimentary of the first phase’s construction of fiber, the fiber network is now extended to areas with the highest demand. In this phase we begin to see the highest capacity (small business and residential users) now being connected directly to the fiber backbone, freeing up the ultra high capacity wireless layer to now deliver more connectivity to the mobility layer while also remaining in place as backup to the fiber.

Concurrently, more ultra high capacity equipment is now put in place so as to add to the next layer down total capacity, upgrading it from hundreds of megabits of throughput to gigabit (or multi gigabit) as needed.

At the edge, distribution is now increased to a shared 100Mbps allowing the residential users the ability to connect (fixed) at speeds burstable to 50 Mbps while still maintaining the WiFi connectivity which allows many mobile users to continue using mobile VoIP/PDA solutions. Depending on market conditions, we may also be looking at deploying WiMAX or potentially another technology, (LTE?) at this time.

With the conclusion of this phase, approximately 15% of all buildings in Philadelphia will now be connected directly to fiber and fiber is now available at every block. This phase is expected to last a total of three years, not including the work done in phase one.

Phase three –

At this time, work begins in earnest to connect every single building to the fiber backbone with the ultra high capacity wireless being kept in place for backup and areas where it would be difficult or prohibitively expensively to install fiber. This would include problem areas such as rivers, railroads, or highways. Upon completion of his phase, at the edge, every single wireless access point is now connected directly to fiber with a wireless connection for redundant backup.

While all this is happening, the originally deployed wireless edge equipment will be upgraded or replaced, depending on demand. By the end of phase three it should be expected that almost all of the existing edge wireless gear will now be experiencing its third complete upgrade. And in order to now service the expected demand, we will also be replacing the layers that feed the edge with state of the art wireless equipment which at that time would probably consist of 10Gbps at the uppermost ring (now dedicated almost exclusively to backup duty) with the multi gigabit radios that used to be utilized in the uppermost ring being redeployed to the next layer down, and the gear removed from that ring finding a use farther down the wireless backbone.

At the conclusion of phase three, three years after its inception, 85% of all buildings in Philadelphia will now be directly connected to fiber and the continued evolution of the wireless cloud will have kept pace with the state of the art.

Phase four –

Now we have hit the home stretch, with only a small percentage of cleaning up to do. However, while the percentage of total numbers is small, these will be the most difficult builds and we have allocated two full years to complete this phase.

While this will conclude the fiber portion of the buildout, all 139 miles of Philadelphia is completely connected to the network by fiber, we are still dependent on a constant upgrading process to keep the mobility segment of this network alive.

There are several aspects of this network proposal that are not illuminated here – nor could they be, For instance, administration, pricing, cost for each phase, adoption rates and the justification for those predictions – but all of that information can be made available.

It’s still all about the money –

So, how does this get paid for? Strangely, this is actually the easy part.

I would point your attention to the following quote, taken from here.

“Atlanta-based EarthLink has given up on its Wi-Fi plans across the country. In Philadelphia, the Internet service provider said it could not find a buyer for the network it spent $17 million to build, and talks to donate the network to the city or a nonprofit organization failed, even after it sweetened the offer with $1 million in cash.”

Hey – here’s a tip – TAKE THE NETWORK AND THE MONEY!
(…and let Earthlink out of their contract, they don’t want the network and nobody wants a law suit)

And this is what you do with it…

First off, use the $1 million to deinstall every single Tropos device and sell it, heck eBay it, for let’s say $1,000/each. I would suggest that the market might not support the $1K/unit price tag so lower it to $500/unit and dump in on any third world country that will take it. Speaking in round numbers, we can estimate that 4,000 Tropos units were put into service in Philadelphia, which at $500/each would provide $2 million to replace that edge equipment with. I could make specific recommendations as to which manufacturer’s equipment I would replace the Tropos gear with but this is not the time or place for that discussion.

Once the network is now operational, the job becomes one of marketing to the business community as well as the residential customers. The idea that Earthlink had only signed up 5,942 customers shows that the service they were providing was not interesting enough to attract any more customers. But, for the sake of this discussion, let’s assume we can double that number to 12,000 customer within six months with an average ARPU (Average Revenue Per User) of $23.00/month. That would now give us $276,000/month in revenue with no debt other than our operational expenses. Year two allows for doubling of the customer base again, with a rise in ARPU to $30.00/month (a goal I would be disgusted if I couldn’t surpass) giving us $720,000/month. Years three and four double the subscriber base while also increasing the ARPU which should ensure a healthy, viable, and sustainable network in perpetuity.

There should also be a component where any service provider can purchase transport or provide services (as in the case of technical support, email, video, etc.) across this resource, and yes, that would include the local telephone company and the cable providers. This would also make a very inexpensive infrastructure for the cell phone providers to increase their footprints with a huge capital expense.

And yes, a full business plan can be submitted should anyone seriously want one.

Now, on a personal note – WAKE UP PHILADELPHIA, you stumbled, you didn’t die. At one point you lead the way in this country and probably created more of a stir than at any time since Ben Franklin’s times.

You have taken the initiative to be the first, you started an entire industry and now have the choice of giving up or moving forward. I believe that given the right justification, as well as a sound plan, your city will be able to lead this country into this new millennium – but you have to want it.

I’m pulling for you, I know you can do it and I believe that deep down you know you can do it too. Now the question is, do you have the leadership to get the job done? If not, don’t worry, there are other cities that will answer that challenge, cities that will be able to make the claim that they did it, that they were forward looking enough to make this infrastructure work and they shall be the one that reaps the yet to be realized benefits of economic development, digital inclusion, and the ability to show the rest of this country how it is done.

Your choice.

John F. Kennedy once said, “Victory has a thousand fathers, and defeat is an orphan.”

Earthlink’s Rolla Huff indicated that EarthLink will stay focused on serving people using
dial-up Internet service and casual Internet surfers who want an economical plan.