I have just returned from an Australian lecture tour, which included meetings with government officials in Canberra. I discovered that Australia is engaged in a vigorous debate about a greater than $36 billion investment in broadband optical fiber. A new, state owned network monopoly, would deliver to every home at least 1 GB of bandwidth connectivity.
The Australian debates prompted me to look into the future of wireless bandwidth to examine the optical fiber vs. 4G wireless tradeoffs.
The International Telecommunication Union defines 4G as a downlink speed of 1 gigabit/sec for stationary or slow moving users and 100 megabits/sec for when devices are traveling at higher speeds.
A 4G network is expected to provide all-IP based broadband to IP telephony, connectivity with laptop computers, access to wireless modems and support of smartphones. Current wireless carriers have nothing like that despite repeated claims of 4G. Technically, what carriers offer are pre-4G, or even 3G and a half capabilities. ITU lets carriers advertise LTE and WiMax Advanced as 4G because these networks are significantly faster than the established 3G technology, which runs at about 14.4 megabits/sec downlink. WiMax can deliver up to 70 Mbps over a 50Km radius.
Then in place the 4G technology will be able to support interactive services like video conferencing (with more than 2 sites simultaneously), high capacity wireless Internet, and other communications needs. The 4G bandwidth would be 100 MHz and data would be transferred at much higher rates. Global mobility would be possible. The networks would be all IP and based on the IPv6 protocol. The antennas will be more capable and offer improved access technologies like OFDM and MC-CDMA (Multi Carrier CDMA). Security features will be significantly improved.
The purpose of 4G technology, based on a global WWWW (world wide wireless web) standard, is to deliver “pervasive networking”, also known as “ubiquitous computing”. The user will be able to simultaneously connect to several wireless access technologies and seamlessly move between them. In 5G, this concept will be further developed into multiple concurrent data transfer paths.
In the United States, the immediate challenge is finding the wireless spectrum. Recent tests in by LightSquared’s ground-and-satellite LTE service found that it interfered with GPS signals. Therefore, the FCC is holding back on proceeding further.
Meanwhile the FCC is considering an interim deployment and operation of a nationwide 4G public safety network, which would allow first responders to communicate between agencies and across geographies, regardless of devices. The FCC released a comprehensive paper, which indicates that 10 MHz of dedicated spectrum currently allocated from the available 700 MHz spectrum for public safety will be used to provide adequate capacity and performance for special communications as well as emergency situations.
While the US is holding back on 4G South Korea has announced plans to spend US$58 billion on 4G and even 5G technologies, with the goal of having installed in S. Korea the highest mobile phone market share after 2012, with hope to set the international standards.
Japan’s NTT-DoCoMo is jointly developing 4G with HP. A the same time Korean companies like Samsung and LG are also proceeding into 4G to gain global market share in advanced smartphones. Recently Japan, China and South Korea have started working on wireless technologies and they plan to set the global standards for 4G.
A 5G family of standards would be implemented around the year of 2020. A new mobile generation has so far appeared every 10th year since the first 1G system was introduced in 1981. The 2G (GSM) system rolled out in 1992. The,3G system, W-CDMA/FOMA, appeared in 2001. The development of 2G and 3G standards took about 10 years from the official start. Accordingly, 4G should start to be installed after 2011. There is no official 5G development project though this may take place only after a 2020 launch.
The development of the peak bit rates offered by cellular systems is hard to predict, since the historical bit rate development has shown little resemblance with the exponential function of time. The data rate increased by a factor 8 from 1G (1.2 kbps) to 2G (9.6 kbps). The peak bit rate increased by a factor 40 from 2G to 3G for mobile users (to 384 kbps), and by a factor of 200 from 2G to 3G for stationary users (2 Mbps). The peak bit rates are expected to increase by a factor 260 from 3G to 4G for mobile users (100 Mbps) and by a factor 500 from 3G to 4G for stationary users (1 Gbps).
Affecting the launch of 4G and 5G wireless technologies is the growth in mobile data traffic at CAGR of 92 percent between 2010 and 2015. Global mobile data traffic will grow three times faster than land based IP traffic from 2010 to 2015. Global mobile data traffic was 1 percent of total IP traffic in 2010, and will be 8 percent of total IP traffic by 2015.
Whether Australia should proceed with digging trenches to every household to secure 1GB connectivity by 2015 is debatable. There are also political considerations.
First, the Australians will have to build a wireless network and erect wireless towers anyway. They will have to provide wireless connectivity for wireless traffic that is growing much faster than land-based traffic. Wireless assets will continue to require steady new capital investments as the S. Koreans, Chinese and Japanese forge ahead with 1GB wireless 4G service prior to 2015 and with 5G after 2020.
Second, the capital costs of any technology upgrades would be always much higher for landlines, based on fiber optic fiber, than on wireless towers. The operations and maintenance costs for last-mile copper circuits that will have to remain place exceed the capital costs of additional wireless towers.
Third, LTE and WiMax Advanced wireless already in place have the capacity to support over 14.4 megabits/sec downlink and up to 70 megabits/sec. I presently pay a premium for a 12 megabits/sec downlink high-speed connection to the Internet. That is completely satisfactory for my extensive Internet usage as well as movie downloads. What the Australian can do with a ready availability of over 14.4 megabits immediately is not clear.
Lastly, there is an issue of how to choose technologies that would lend themselves to market competition vs. monopoly operations. The capital cost for placing fiber optic cable is an investment that has a technology life of over 50 years. Fiber optic cable to every home suits a monopoly model of pricing, since the costs of the delivery of services are unrelated to the costs of capital investment. In contrast, once wireless towers are in place, they can host a variety of wireless providers at costs that can be adjusted to reflect geographic characteristics. There is no question that proceeding with wireless connectivity would allow diverse competitive offerings to co-exist.
With the prospects of a rapid evolution in wireless connectivity as well in consideration of the widespread dispersal of Internet users in Australia, it does not appear to make sense to commit immediately to expensive fiber optic circuits to support to deliver Internet services to the Australians.
Paul Strassmann is the Distinguished Professor of Information Sciences, George Mason School of Information Technology and Engineering and researcher at the Centre for Secure Information Systems. He is also former CIO of NASA.