Who benefits from network designers' choice of network architecture? When they make a fully informed decision, they do. A fully informed decision can result in the "holy grail" of networks. This ideal architecture results in both lowest initial installed cost and lowest life-cycle cost. In most cases, this architecture will be either fiber-to-the-desk (FTTD) or fiber-to-the-zone (FTTZ).

I intend to make the following three points:

• The choice of network architecture has a strong influence on the network cost.
• The parties involved may not benefit from network designers' best choice.
• The cost impact of your choice can be hundreds of dollars per node more than another option.

In a new build or a building rehabilitation, network designers have three network architecture options:

• Traditional horizontal unshielded twisted pair (UTP) with vertical fiber
• FTTD
• FTTZ/fiber-to-the-enclosure

All three comply with current standards.

If network designers choose the first option—hereafter called the UTP option—they will incur increased initial installation costs and increased life-cycle costs. Increased initial installation costs will result from the costs associated with the telecommunication room (TR), which is required on each floor of a multi-story building and in each wing of a multi-wing building. The cost ranges from $20,000 to $50,000 per room. This range was developed by network managers and designers who pay for TRs. If each TR supports 48 nodes, the TR cost per node will be $416 to $1,041. In addition, network managers will have to purchase more switch ports, as port utilization will be at roughly 70 percent, according to Corning Inc. Finally, network managers will have to install a UTP network now and another one in the future. History suggests that reinstallation runs $75 to $150 per drop. I will use $100 per drop for convenience. In summary, the UTP option results in the following:

• A TR cost of $416 to $1,041 per drop
• More switch ports
• A future upgrade cost of $100 per drop

If network designers choose the second option—the FTTD option—the TR cost of $416 to $1,041 per drop will be eliminated. In addition, network managers will purchase fewer switches, as port utilization will be high—roughly 90 percent. Again, this is Corning Inc.'s number. Finally, the fiber network will last much longer than a UTP network. History has shown this to be true. For example, a 50 ?m fiber installed in 1978 meets the requirements of gigabit Ethernet (1998) and some 10-gigabit Ethernet links. With UTP, each upgrade in bit rate requires new cabling.

On the other hand, the fiber media converters will cost $100 to $450 per link, depending on the bit-rate requirement. In summary, the FTTD option results in the following:

• Elimination of TR cost of $416 to $1,041 per node
• Fewer switches
• Elimination of a future upgrade cost of $100 per node
• Increased cost of $100 to $450 per node

This simple analysis shows that FTTD has the potential to reduce network cost by hundreds of dollars per node. A more detailed analysis shows that savings can range from $20 per node to $189 per node. For a moderate-size network of 500 nodes, this range means a savings of $10,000 to $94,500. I expect that most readers will consider this last number significant.

If network designers choose the third option—FTTZ or FTTE—they will eliminate the TR cost, reduce the number of switches, and eliminate future upgrade costs. In addition, they will reduce the media conversion cost per node by the number of ports served by each zone switch. In summary, the FTTZ option results in the following:

• Elimination of the TR cost of $416 to $1,041 per drop
• Fewer switches
• Elimination of a future upgrade cost of $100 per drop
• Increased cost of $4 to $18.75 per node (assuming 24 ports per zone switch)

Who will benefit? If network designers choose the UTP option, the switch manufacturer, the UTP manufacturers, and the installers will benefit. The switch manufacturer will sell more switches. The UTP manufacturers will sell products now and again in about five years. The installer will sell services now and again in about five years.

What is the cost impact? These savings can be significant. A detailed analysis of 12 scenarios indicates a savings range of $19.83 to $521.28 per node. More impressive is the fact that the comparison is biased against fiber by $93.75 per node! If I remove this bias, the reduction becomes an impressive $615.03 per node!

If network designers make a decision based on incomplete or inaccurate information, product and service vendors benefit, as presented in the previous paragraphs. In addition, more is spent on the network than is necessary. Information from these vendors will, by necessity, be biased (no criticism is implied here).

In making a decision, chief financial officers or chief informational officers must use information that is unbiased and represents the whole picture, not just a piece of the picture. A decision based on initial total installed cost is usually the best decision.

Educational Content Provided by Pearson Technologies