Robert Liu verifies early speculation that the Philly network needs indoor bridges: A CPE (customer premises equipment) could be a T1 modem, a cable modem, or a DSL modem--or, in the case of many Wi-Fi mesh networks, a Wi-Fi to Ethernet bridge with a high-gain antenna. This has been the missing piece in much of the coverage of Wi-Fi mesh networks.
Critics have said that cards built into laptops or added to desktop machines couldn't receive signals from ubiquitous outdoor metropolitan networks. They're generally correct. But until this fact-finding pre-proposal meeting in Philadelphia that Liu attended at which potential vendors had a chance to ask questions for the first time it was unclear that a CPE was required. The RFP states on page 10:
Should additional customer premise equipment ("CPEs") be required or assumed in order to deliver this in-building coverage, Respondents are expected to state this in their Proposals and elaborate on this requirement and their assumptions.
Liu says this is now a given, not a "should."
I disagree on this adding substantial cost, however, as Liu writes: a Senao 200 mW 802.11b bridge retails for under $100; in quantity, it would be substantially less. Because the network builder will be a wholesaler, they can't recover the cost of this device from the consumer directly, but it appears they'll be responsible for CPE selection. This could be built into the price of the network, representing a couple dollars per month of the wholesale cost. Or they could offset the cost to ISPs who recover it from end-user fees or leases. The Senao is being used in some Tropos deployments right now as a CPE.
Liu reports as well that the information provided to vendors who want to bid is probably inadequate for true RF planning: building footprints date to 1996; no building heights were provided. Other topographical details about poles and other city facilities was available, however.
CPE is not really the killer in these details, it just increases the payback period on subscriber acquisition. And CPE should be a given, because in the wild urban environment, you can't control what your apartment neighbor puts on the other side of your wall. Heck, we need CPE out on Fire Island! Imagine the Tier 1 support nightmare that will ensue when somebody plops an interfering AP with a 12 dBi omni attached to their VZ DSL out there in the middle of 10 Wireless Philly subscribers. If any of you bidders are thinking about using 802.11b/g as the sole demarc between provider and subscriber, you're out of your heads. And to make this edge technology the mandatory way to connect, the RFP writer would be nuts too. You could build it for the projected $10MM price, but you wouldn't be able to manage it for very long.
The real killer of wi-fi and the cost spiker in this proposal is the requirement for "seamless, in-motion usage throughout the Coverage area by subscribers with service provisioned through a SP. This includes the ability for subscribers to maintain "session-level" persistence while the subscriber's device is in motion at speeds up to sixty miles per hour." This capability MUST (caps mine) be supported with no interruption to applications running on the device." Page 8
If you have to build out a network with this capability, with current wi-fi technology, in an urban environment, I don't see it being done and meeting an acceptance test to get paid, (all due respect to RoamAD and Wi-VOD, Rio Rico to Green Valley in Arizona is not the same RF environment as Philly, the proximity of Fort Huachaca notwithstanding)
If you drew a line in the sand today, or say, on the 23rd of May, and be required to build it with technology available on that day to meet the spec by the 23rd of May 2006, you would/will? have to be either an MVNO or a spectrum owner/lessor.
Plus, what happens when the city doesn't own the real estate necessary to mount the equipment to fill a dead spot?
The security standards section ( not all of the requirements in this section are security standards ) also dictates the provision of a smart-edged network, supporting multiple-SSIDs and VPNs, jacking up the cost.
One of the great things about this proposal is the City being the anchor tenant for it, providing a base revenue level, the other thing is the tiers of service, fixed, nomadic, and mobile. The "triple play" of data services, so to speak. It provides the distinct possibility of success. The only way you can get this now in New York, is to get VZ DSL, VZW EV-DO and use the VZ Wi-Fi payphone hotspots.
The winner of this deal should be somebody that takes the RFP as the base document and crafts a network design that can deliver the services described in the Business Plan document, which is where the real meat is. It should be a counter-proposal that uses Wi-Fi where appropriate, and other licensed and unlicensed technologies as they fit. If it's a Wi-Fi only edged net, it will be doomed to failure.
The other great thing about this proposal is its being brought forth at all. The technological growth of this country is being strangled by a combination of the incumbents turf defense and the ownership of our state and federal governments by the large corporations. It is somehow fitting that Philadelphia should be the issuer of this RFP, a broadband Declaration of Independence from the tyranny of the telcos. Just remember, along the path leading to the Declaration and up to the Constitution, there was a bloody Revolution. I think this certifies we're in another. Let us rid ourselves of the oligarchs!
Here it comes. The typical government project, ill defined, open ended, cost over runs, and ultimately does not do what everyone originally believed it would do. Then come the inevitable years of fingerpointing, arguing, lawsuits. Tell me again why this is a Good Idea?
[Editor's note: I think you should read Philadelphia's detailed business plan and RFP before making that statement. It's interesting that it started with CPEs being optional and moved quickly to mandatory. But that seemed more like flexibility for the vendor to choose technology.--gf]
Obviously the editor and I disagree on the value, let alone the necessity of government involvement in this kind of enterprise. I am reading the Philadephia business plan, and I am struck by the numbers of assumptions which I and more importantly which many scholars believe are fallacious. For example, the BP talks about cities being the great creative engiones of commerce, culture and society throughout history, and mentions cities' central roles in the railroads and the automobile. However, modern objective reasearchers note that government involvement in these industries led to massive waste, fraud, abuse, and poor workmanship. See chapter 8 of the Politically Incorrect Guide to American History by Thomas Woods, and read the research of Burton Folsom, for example.
There is no evidence at all that government subsidy of anything results in something better for the economy in general. If it does, then that serves only as proof that lower if not elimination of taxation of enterprises is the real cause, not government involvement.
I note also that Philadelphia will create a "nonprofit corporation" to oversee this wireless project. More bureaucracy, more hearings, more places for lobbyists to lurk, and in the end more overhead that could be better devoted to what the free market can provide.
The argument here continues to be that of government control of enterprise versus letting free enterprise operate without having to deal with the idiocy of government. I believe government has no place in most enterprises. That means garbage collection, street cleaning, electricity production and distribution, telephone service, automobile manufacture, internet service provision, and so on down the line. Can the editor make a convincing case for government involvement in any of the aforementioned enterprises? Which ones? All of them? Why or why not?