“If it were just about cell phones, we wouldn’t need 5G,” says Verizon CEO Hans Vestberg. In other words, 5G is about new use cases and applications other than supporting smartphones.
Maybe that will ultimately not be the point. If you remember the major revision of U.S. telecommunications law called the Telecommunications Act of 1996, you will remember the practical consequences of deregulating the local telecom access business.
Revising U.S. law, the Act enabled competition for local telecom services, lawful operation and ownership of Class 5 voice switches, the right to sell customers voice and other services and wholesale access to incumbent networks.
All that happened just prior to voice communications reaching a historic peak about 2000, with a rapid decline. Most incumbent telcos lost 35 percent of their customers for that service in 10 years, as much as 65 to 70 percent over two decades.
Service providers also lost half their revenue from long distance calling over that same period.
At the same time, other big changes in end user demand were happening: substitution of mobile phone service for fixed service; use of mobiles instead of cameras or music players, GPS devices or video screens.
There also was increasing use of the internet as a substitute for a wide range of other activities and products. In 1996, for example, it is estimated there were 36 million global users of the internet, representing less than one percent of the world population. A decade later, that had grown to 17 percent.
About that time, some 14 percent of the U.S. population was using the internet, on dial-up connections. A decade later, that had grown to about 66 percent.
The point is that disruptive changes in regulatory framework can produce outcomes we did not expect, especially when disruptive enabling technologies happen at the same time, allowing massive product substitution and behavioral changes.
The same thing might happen with 5G. It arrives in tandem with other key technologies and platforms, including commercial artificial intelligence, edge computing and internet of things. It may, in the end, be hard to separate the various threads from each other.
In part, that is because computing architectures shift over time, oscillating between centralized and decentralized approaches. That puts computing resources at different places within the architecture, fluctuating between centralized and decentralized designs.
In the mainframe era, computing resources were centralized at far ends of the network. That shifted in the client-server era to more local processing on devices themselves or on local servers. In the internet era computing switched back to far end hyperscale data centers.
But most observers believe we are now in a stage of shifting more workloads back locally, to take advantage of artificial intelligence, heavy local analysis of sensor data to support the internet of things and compute-intensive applications using virtual or augmented reality.
“These days lots of companies want to turn bandwidth problems into compute problems because it’s often hard to add more bandwidth and easier to add more compute,” said Andrew Page, NVIDIA media group director of advanced products.
No comments:
Post a Comment