Digital wireless technologies were designed to answer issues of quality, capacity and cost. But by releasing huge swaths of new spectrum, the Federal Communications Commission has changed the rules, and capacity is now much less important than cost-but not everybody in the industry has realized this.
This question and answer set is intended to correct some common misconceptions about the U.S. personal communications services market and the impact it has on technical issues in the economics of the industry.
Q. How will the U.S. wireless market evolve?
A. Already, one in eight people in the United States has a cellular telephone. By the year 2000, the number will be around one in three. In another five years, it will be two in every five. Soon after, the number will climb above one per household. Based on startup timing, I extrapolate that these will likely be split:
As shown in Figure 1a, by the year 2000, cellular penetration will double to 26 percent (a substantial slowing of today’s growth rate to an average of 15 percent). PCS penetration will surge from near-zero in 1996 to about 4 percent of the population in 2000.
By the year 2005, cellular will have grown to 34 percent and PCS to 10 percent. Those are average growth rates of 5.5 percent and 20 percent respectively.
As shown in Figure 1b, by 2000, A- and B-block PCS operators will have 3 percent penetration divided about equally between them. The C-block carriers will be at perhaps 1 percentage point, and D-, E- and F-block firms will divide about a percentage point among the three of them.
By the year 2005, the A- and B-block carriers will have 5 percent to 6 percent divided between them, the C-block operators will have just about caught up with about 2.5 percent, and the D-, E-and F-block companies will have perhaps 2.5 percent split among them.
Q. Will PCS systems be coverage or capacity driven?
A. They will be coverage driven for the next decade. To illustrate, I will choose the Chicago basic trading area as a model, and Code Division Multiple Access technology, since it is the majority choice in the United States.
It takes about 210 three-sector CDMA base stations to provide full coverage to the approximately 20,000 square kilometers of the Chicago BTA. These cells will range from less than a kilometer radium in the inner city to several kilometers in the countryside.
The capacity of a three-sector CDMA cell with a 30-megahertz spectrum allocation (11 radio transceivers per sector) is about 400 Erlangs, including 40 percent overhead for soft handoff. So the capacity of this 210-cell system is about 84,000 Erlangs. With only 10 megahertz of spectrum, that number falls to about 21,000 Erlangs. An A or B operator in Chicago may have about 135,000 customers in the BTA section of the market by year end 2000. Then the total offered traffic in the busiest time of the day might be 0.03 Erlangs per customer times 135,000 or 4,050 Erlangs.
The capacity of a basic coverage system is underused by a factor of more than 20. In 2005, the underuse will have fallen to a factor of 10. On average, these base stations still need barely two radio transceivers per sector.
Using the same math and logic, we can conclude the C operator has an underuse factor of 70 in 2000, and 11 in 2005.
The D/E/F players have only 10 megahertz of spectrum, so they have less capacity (three transceivers per sector). In 2000, I see them with an underuse factor of more than 50, and in 2005 eight. One radio per sector is far more than is needed in all cases.
I have performed similar calculations for Global System for Mobile Communications and Time Division Multiple Access Interim Standard 136 technologies and conclude that none of the PCS systems will run into capacity in the next decade.
Figure 2 shows how the basic CDMA PCS coverage layout of cells is overbuilt in the capacity sense for the coming decade.
Note that beyond the coverage cell scenario, there is headroom of another 100-fold by shrinking every cell to one-10th of its coverage size. I conclude that:
Except for traffic “hot-spots,” no PCS operator will have any significant capacity-driven cell buildout in the next decade. The capital costs of PCS will be heavily front-loaded.
10 megahertz players will not be significantly penalized for having less spectrum than 30 megahertz players. This is reflected in the FCC’s current consideration of rules to permit the private sale of “surplus spectrum” by licensees.
The vendors have come up with the wrong products for PCS in the USA. The industry needs maximum range, to minimize the number of base stations. There is no requirement for high capacity, the driver the vendors have been addressing.
Q. As a PCS player, how can I enhance my profitability?
A. Revenue enhancement. Soon “everybody” will have a portable phone. The current downward usage trend will reverse and primary use of wireless telephony will grow. But this is no cause for complacency. Growing the market will take a lot of hard work and “out of the box” thinking. Some examples:
Primary use. Wireless phones will undergo a transition from secondary to primary use once the industry has addressed quality, privacy and cost issues.
Data and messaging. One industry source says that data will be only 4 percent of all wireless traffic in 2000, growing to 7 percent in 2005. Finding ways to grow data services will be a critical factor for success in the coming market.
Private wireless services and public/private integration. The 1.9 GHz unlicensed band is no real substitute for “clean” spectrum. Business customers will be prepared to pay for services that are interference free and can be more readily integrated with the public services they also buy. PCS operators will have surplus spectrum on which to provide these services.
Capital cost reduction. The largest capital cost elements are proportional to the number of cell sites that have to be built. There are two parallel approaches to controlling these costs.
Reduce cell site count by using various technical approaches.
Share cell-site and other facilities with other operators. Some local authorities are either mandating, or at least encouraging, the use of shared facilities. Shared backhaul and switching are other capital-saving possibilities. Sharing helps with both capital and expense costs.
Operating cost reduction. Outsource as many activities as possible. Many of the functions that are now internal become uneconomic when recreated in multiple operators’ organizations. Horizontal integration of the industry will be critical to cost reduction.
Get the costs out of selling. Today’s $30 bubble-pack phone is activated by a phone call and billed to the customer’s credit card. This approach saves hundreds of dollars per added subscriber, and ongoing savings in billing and collections.
Reduce cell site count. Maintenance, taxes and lease costs are charged per site.
Note that “cell site cost reduction” appears in both the capital and expense sides of the equation. This is the most influential cost element, and millions can be saved through the use of straightforward measures.
Q. How much does a cell site cost?
A. Fixed costs for a base station include capital and ongoing costs. Capital costs include real estate acquisition, permissions, zoning, public hearings, engineering design, spectrum clearing and coordination construction and non-traffic driven electronics like power supplies and backhaul.
Ongoing expenses include lease payments for the site, routine and emergency maintenance, real estate taxes, permits and usage fees, utility costs and backhaul facility leases (if not capitalized). Adding together capital and expense, the equivalent impact of every stand-alone base station on the NPV of business is in the region of $550,000 to $700,000 in fixed costs. Cost per Erlang for network equipment is extra.
I have extrapolated these costs onto
market predictions for 2005 for GSM and CDMA technology, and divided them on a “per subscriber” basis. At full capacity, a 30 megahertz system would cost $195 per subscriber, while a GSM system would cost $321 per subscriber. A, B and C operators would pay $600 per subscriber using a CDMA network and $925 for a GSM network. D, E and F operators would spend 2,000 per subscriber using a CDMA air interface and $2,800 per subscriber using a GSM air interface. These figures include all network costs, but do not include the cost of spectrum.
For comparison, the cellular industry has spent some $740 in capital costs per subscriber to date, and have incurred zero spectrum costs. So all PCS operators have strong incentives to save money on their buildout.
Q. How can I save money on cell site costs?
A: Getting a better price from the equipment vendors, though always a good idea, is not the most impactful way to save money. It is more practical to halve the fixed costs of cell sites. It is very clear that, in systems that are not capacity-limited, this approach has a larger payoff than attacking equipment costs.
Q. How can I build fewer cell sites?
A: To do this, cells must be twice as large in area, or equivalently 41 percent larger in range or radium, meaning we must increase the link budget. In typical regions the required improvement varies from 4.5 dB in low-density suburban or rural country to 6.8 dB in high-density urban areas. Unfortunately, at 1.9 GHz, link budgets already are disadvantaged compared to cellular incumbents at 850 MHz by about 7dB if they both use the same air interface technology.
There are several ways to accomplish the goal. Some, like increasing tower height, are simple and obvious, but impractical in today’s zoning environment. Others are more practical:
Use tower top electronics, including cryopackaged systems; and use “smart antenna” systems.
These approaches are applicable in combination to the same system-they are not mutually exclusive.
Use tower top electronics to improve the return link budget. In PCS, it is becoming common for operators to use tower top electronics to improve the link budget by avoiding tower transmission line losses. For conventional tower-top; packages the improvement can be on the order of 2 to 3 dB, resulting in a 22 percent to 31 percent decrease in the number of cell sites required.
A greater improvement can be obtained using cryopackaged tower top electronics with a superconducting, low-loss, high-performance filter and a cryocooled low noise amplifier. Cryopackaged electronics of one kind or another are being manufactured by SCT, STI, Conductus and Illinois Superconductor.
Use “smart antenna” systems. Most antennas for PCS three-sector cells will have gains in the region of 16dB, or 40 times. An array of 18 20-degree smart antennas may have as much as 4 or 5dB of additional gain, counting the additional losses from the switching matrix. It is sufficient to reduce the cell count of a typical system by as much as 57 percent. Smart antennas are under development by a number of large and small manufacturers, including ArrayCom, CellWave, MetaWave and Hazeltine.
I strongly believe that the PCS industry could be in peril if it ignores the above issues. In particular, the failure of the vendors to recognize these issues is curious, since more and more PCS deals are being struck that involve vendor financing of the operator’s equipment. The vendors now have direct interest in ensuring that they equip their customers as cost efficiently as possible. If they do not, they will wind up holding a lot of worthless paper. It remains to be seen whether or not they are paying attention.
Dr. Ray Nettleton, president of NettWork Consulting L.L.C., which provides consulting services to the wireless telecommunications industry, including Superconducting Core Technologies Inc. and UniSite.
