Leading Network Transformation
Wireless Networks: Starting Over?
If you were to start all over again, what would you do differently?
Well, my answer is many things. For example, I would have stopped at the yellow light a few months ago, rather than speed through it, which resulted in a citation. I would have left a day earlier for my best friend’s wedding because the flight delay caused me to miss part of the wedding. I would love to have taken a few art history classes while in college, and so on. Looking back, it is easy for many of us to see opportunities to change course to improve upon an outcome. Looking forward, it is much more difficult to make decisions without knowing the absolute outcome.
I recently participated in the Telecom Council Panel in Silicon Valley. One of the questions was if we were to start all over again on the wireless network, what would we do differently? Again, the answer is many things.
Before answering the question, let’s take a walk down the memory lane. Wireless communication can be traced back to the ancient time of smoke signals. However, for our discussion, let’s begin with AMPS (Advanced Mobile Phone System), also known as 1G. AMPS began in the early 1980s. The idea of a phone not attached to the wall that can be carried around is intriguing. Having the ability to call while on the go is worth something to many people. However, the phones were bulky and expensive (~$3000 for “The Brick” and ~$1000 for “The Bag”). The network was not built out and service was spotty. In addition to the $50 access fee per month, each minute cost $0.99, and if you leave the local area, each minute would cost $2.99 because it is considered “roaming.” Remember, this was during a time when a new Honda Civic cost around $5000.
A Niche Market
Due to the cost, many believed that cellular telephony was a niche market. ATT, along with McKinsey Consulting, published a report concluding that the TAM (Total Addressable Market) for the US would be around 900,000 subscribers by year 2000
In the early 1990s, digital technologies known as 2G (GSM, TDMA, CDMA) started to emerge making cellular telephony more efficient. The digital network equipment and phones were smaller and less expensive. For the same capital, network operators were able to build more cell sites to cover more territory. The less expensive phones attracted more subscribers. There were also new entrants to compete with the duopoly by driving down service cost. This economic flywheel effect proved the McKinsey report wrong as it grew the wireless customer base in the US to 109M by year 2000.
The Internet Age
In the early 2000s, the Internet was booming. Similar to the desire for mobile voice communications, there was strong consumer demand for mobile Internet. Therefore, 3G (EV DO, HSPA) emerged to provide data services in the low single-digit Mb/s range. With the advent of smartphones and more content, such as videos and large files being consumed on the go, 3G speed was good─but the experience could be better.
To handle the tremendous growth of data traffic, WiMax was launched by one operator and touted as 4G circa 2010. Since there was only one operator with WiMax, it lacked the economy of scale and failed, even though the technology was good. Alternatively, other network operators deployed LTE as 4G networks. In addition to being more efficient, LTE offered better performance, and the purchase volumes lowered the capital cost significantly for the network operators. Once again, economics fueled the network expansion to support usage and customer growth.
A Great Success
By all measures, the wireless industry has been a great success in the last three decades. This success is due to the mobility that consumers deem critical for their communications. More importantly, it grew from a niche to a mass market due solely to the economics. The wireless industry would not be what it is today if the cost of services, and the equipment to provide the services, remained high.
Remodeling the Past
So let’s explore what we could do differently in the wireless network to make the industry even more successful than it is today:
First is the radio access technology. AMPS was a common standard but the divisions between GSM, TDMA, HSPA, CDMA and EVDO as 2G and 3G technologies caused industry fragmentations. So if we start from scratch, we should have one unifying radio access standard.
Second is the hardware. Since the growth in the future will be in connecting Internet of Things rather than phones, the existing network is insufficient because it cannot economically scale to support the billions of connections. The solution to this problem is to avoid using custom-made hardware where rapid growth requires “fork-lifts.” Rather, we need efficient software running on COTS hardware that can massively scale. This is also known as virtualization.
Third is having a flexible network that can support all services. There should not be a voice network and a separate data network.
At the very least, it needs to have the following capabilities if we start from a clean slate:
The air interface needs to operate at the Shannon limit with a robust interference- tolerant coding scheme
It needs to provide high definition multimedia services
It needs to provide broadband speed at least a magnitude of order faster than what LTE can deliver today
It needs to be highly responsive to support services such as UAV (Unmanned Aerial Vehicles) or self-driving cars
It needs to scale to support billions of IoT connections
5G is Coming!
As the industry is defining 5G, these are some of the requirements we have in mind. However, the economics have to work for 5G to continue growing at a fast pace. Economics made it work from 1G to 4G because revenue and traffic grew at a similar pace. The industry is now at a critical juncture because traffic is growing around 50% annually, yet revenue growth is growing around 3%. The success of 5G will depend on changing this course. Therefore, 5G cannot be just a niche market. Otherwise, it will take off the way the Concorde did for the airline industry.