Moore states that at a constant cost, the amount of computing capability doubles every 18 months.
Oldefar's corollary states that number of units supported by a given IT support position increases at twice the rate of computing capability.
I will be honest - my corollary is based on observation since 1972, and by an intuitive feel for the ratio rather than scientific study. Maybe I will get ambitious and plug verified numbers in for various support roles.
In 1972, computing was mainframe, communications was primarily analog on either dedicated lines or using accustical couplers over dialup. Hardware was often a mix of tubes and discreet solid state devices with a rare IC in some applications. Backplanes were hardwired, either soldered or wire wrap. Since that time I have seen a variety of highly prized skills become nearly obsolete - high value soldering and micro-circuitry repair, oscope data analysis, machine language fault isolation, RF alignment, cable bundle splicing, RF cabling, fiber optic field repair, analog circuit balancing, telephone istrument refurbishment and repair, cable engineer. Among support roles that have been or are being absorbed into the common user role include typist, word processor, presentation specialist, CAD-CAM specialist, Web developer. All these roles still exist, but they have not expanded with the growth in connections. A typical office admin handles the presentations and Web support. Users do their own typing, word processing, and CAD-CAM.
In 1984 and 1985 I led a 20 man communications team responsible for the installation, maintenance, and repair of about 300 data and voice T1 circuits, a couple of private DS3 systems, and the associated channel extensions and tail circuits, a host of dialup connections, and several satellite links. An additional 10 technicians supported the controllers and terminals. Our coverage area was a corporate campus and several sites in the surrounding area. An additional 12 people provided operational support - initial alarm notification and problem assignment. Order processing, circuit and network engineering, cable engineering, and project management employed an additional 1,000 individuals nationwide.
By 1994, that 12 man team had grown to 30, but the scope had expanded from a portion of a state to world wide. The number of circuits supported had increased over 100 fold. The number of operators per shift had dropped to 8. Support expanded to all data, voice, and video systems for what was then one of the largest private networks in the world. Field or onsite support worldwide had dropped from a high of 2400 to about 200.
In 2003, my son supports a 7 state region with an additional 100 fold increase in capacity as measured by bps and end points. He also has PM and engineering responsibility for the area. The shared support team is about 100. The operations team nationwide is 7 per shift.
My point is that despite a doubling of connections every 10 months, the growth in IT penetration is being exceeded by the inherent capabilities of the hardware and software. The result is that the growth in devices is not generating a growth in support positions any longer. The trend will eventually level off to some support level that will remain constant as IT systems expand. At that point, the remaining positions will be filled to handle attrition primarily. While at any given time there will be a short demand spike in a given area, the spikes will be transitory and not growth.