Most technology nerds know “tel” as a prefix meaning “transmission over a distance,” as in telecommunications, television, or telemarketing. Most are unfamiliar with an altogether different meaning as found in the phrase “technology tel,” which is the modern and digital equivalent to an archaeologist’s tel.
Archaeologists define tel as a mound created by many generations rebuilding on the same location, where past efforts become foundations for new generations. This meaning of tel arises in locations that humans have inhabited for many centuries, such as the ancient cities of the Near East and Mediterranean.
Similarly, a technology tel accumulates generations of digital infrastructure built in one location, where it serves interdependent or independent functions. Technology tels exist in every major city, and though city dwellers interact with them regularly, they tend to blend into the background.
Unlike its archaeological counterpart, a technology tel accumulates material in a matter of decades or faster, which makes it possible to observe and illustrate its economic determinants. Those economics occasionally gain attention from developers and city planners, and they are worth some attention from anyone interested in the modern digital experience.
Technology tels grow due to three economic factors:
• infrastructure is built in locations where the most users create the most demand,
• the presence of a prior generation lowers the cost of installing the next, and
• technical factors raise the productivity of collocating two distinct pieces of equipment.
These factors reinforce one another, keeping equipment in the same location.
Southern Manhattan contains the mother of all technology tels. Long before the era of modern technology, anyone affiliated with financial markets already crowded into the district. Then, electronics worked their way into the location to serve the industry. It is obvious why. Financial users have long been lead users of information technology, so every financial office needed equipment from the frontier. Accordingly, after the early 1980s, the district became home to LANs, laptops, servers, screens, printers, and WiFi routers. Plenty of advanced infrastructure also found its way there, both to serve the same set of users and to increase effectiveness through collocation. Many workers use the same ISPs, Internet cables, 4G antennae, LAN maintenance staff, and hosting facilities.
Although sharing is efficient, organic growth is uncoordinated and does not stress efficiency. Instead, competition induces duplicate supply. Multiple upgrades of buildings attract the best tenants, so competitive incentives induce earlier upgrades, and the cycle continues—keeping the technology tel near the frontier.
Users and equipment compete for space, driving up its price. And so arises one of the most telling symptoms of a crowded technology tel: Despite expensive real estate, some buildings house data-centers filled with servers for nearby users who demand low latency. Those datacenters also hap-pen to be in place where there is easy access to the fiber. That is as it should be. The functional value of the datacenter is highest there, which contributes to the willingness to pay for the space.
Today, crowding is visible everywhere. Data equipment, servers, and switches inhabit every closet and basement, and antennae and dishes crown the top of every building. Data lines snake through every tunnel and available duct, including Manhattan’s bridges. Wires traipse every elevated path-way, especially telephone poles, if available. Aesthetics are optional. Equipment goes wherever it can fit and, in some places, where it should not have gone.
Most crowded technology tels generate spawning, that is, the movement of activity to other locations away from the technology tel, either to escape the price of space or the lack of available space. For example, today, the equipment for the electronic New York Stock Exchange and related ex-changes for high-frequency traders sits in northern New Jersey. These are organized around some datacenters, which, if you think about it, are just miniature technology tels coordinated by a manager.
Crowding and spawning are not unique to southern Manhattan. The same economics created other technology tels in many other locations. Consider the second-largest financial center in North America: Chicago. Yet again, that city contains a large network of preexisting tunnels and conduits to support an enormous knot of wires and connections. All the electronics inside the down-town loop today take advantage of it. In another parallel, part of the electronic infrastructure to support the electronic exchanges also left the financial district. Part resides at a datacenter in Arora, a western suburb, and part resides at a datacenter at 350 E. Cermak, two miles south of downtown.
The latter is a retrofitted printing building (which used to print the Sears Roebuck catalogue, among other things). It is a massive building containing nearly infinite conduits and pipes. With more than a million square feet, I see little reason to doubt its claim as one of the largest datacenters anywhere.
The accumulated investments of a technology tel create a number of opportunities and headaches. Catalogue them as growth, neglect, and security.
Growth looks like the attractive part of a technology tel. Once started, an existing location acts as a magnet for more investment and people. For example, in the last few decades, the technology tel built around San Francisco’s financial district spawned a nearby technology tel in the area south of Market Street—albeit, with some help from city planners. That area was among the seediest parts of the city during my childhood, but now it’s a showcase for technology-led urban growth.
As in the case of San Francisco, organic growth often does not just happen on its own. That motivates a question: Can growth be jumpstarted in a new location by a city planner? Some cities in low-density locations have tried to do so through abatements for property and/or sales taxes, as well as expedited permitting. In recent times, this has worked with a propitious situation, namely, a location next to (transcontinental) access lines and inexpensive (overbuilt) electrical supply.
Critics of such programs note that these programs mostly attract datacenters, which create limited employment for construction work and modest employment thereafter. A moderately sized datacenter employs a few dozen people—maintenance and security staff, and a few technicians—and has limited local consequences. It mostly acts as a footloose secondary facility for backup storage for distant users. Giant warehouses and low-rise buildings (with little architectural beauty) typically do not result in a particularly attractive business district.
The addition of a few steady jobs hardly seems worth the effort and expense. So why do so? For one, if enough such buildings can be located in the same area, it can motivate investment in shared infrastructure, such as cellular towers, which spurs further investment. For two, these mini technology tels generate little traffic, meaning there will be few burdens for public administrators. Many planners in low-density locations would prefer that over nothing.
Even with such modest payoff, it is easier said than done. A thriving technology tel depends on each piece working well. Ask a city planner, and he or she will explain the challenges of lining up all the pieces. A local electrical, cable, or cellular company might fail to keep up with the frontier; a freeway exit might need repaving; or a state agency might drag its feet issuing permits.
Some difficult economic problems also can get in the way. For example, the deployment of 5G has many city planners vexed. The prior upgrade between 3G and 4G lowered the costs of deployment by using the existing rights-of-way and infrastructure, such as towers and pathways for backhaul of signals. As of this writing, however, 5G signals do not operate like 4G signals. That could drive up expenses in a local technology tel that wants to remain relevant.
Security risks raise different issues. Consider the technology tel in Ashburn, Virginia, next to Washington Dulles International Airport. That location became focal because it contained one of the original switches for the Internet, and then AOL and other dot-com wunderkinds moved in. For more than two decades, it has had a momentum all its own, with almost no spawning. Today, this location contains an enormous collection of datacenters, Internet switches, and private firms who take advantage of the connectivity, making it one of the largest technology tels in North America. Any firm trying to service the Eastern Seaboard or the federal government invests there.
Notice the historical irony. Building secure communication infrastructure in military applications provided one of several motives for DARPA to invent packet switching many decades ago. Yet, economic forces have created just about the opposite of what the military wanted: potential for a single point of technical failure due to a concentrated geographic target.
Interdependence links one failure to another in a technology tel, and this too creates potential head-aches. One of Chicago’s experiences illustrates this well. In April 1992, some misplaced pile drivers in the Chicago River started a leak in a conduit tunnel, which was first discovered by a workman inspecting fiber lines running through the tunnel. The city ignored the warnings, and the unrepaired leak eventually became a flood. Many basements were served by interconnecting underground tunnels, so the flood spread. Swimming fish appeared in basements. It took a while to clean up. Settling the insurance claims was an even bigger mess.
A technology tel is propelled by the relentless urge to build and agglomerate. Crowding and spawning have become part of the landscape. Take a look around you. It is everywhere and notice-able, if you just look for it.
Copyright held by IEEE. To view the original essay click here.