- 1.1 The Frontier: Millimeter Wave Wireless
- 1.2 A Preview of MmWave Implementation Challenges
- 1.3 Emerging Applications of MmWave Communications
- 1.4 Contributions of This Textbook
- 1.5 Outline of This Textbook
- 1.6 Symbols and Common Definitions
- 1.7 Chapter Summary
This chapter is from the book
This chapter is from the book
This chapter is from the book
1.3 Emerging Applications of MmWave Communications
60 GHz WPAN and WLAN are only the first step in a mmWave communications revolution. In addition to providing the first mass-market mmWave devices and enhancing cross-disciplinary communications design, 60 GHz communications will also have a substantial impact on other network technologies. Data centers may cut costs by employing mmWave communication links to interconnect the computers for high bandwidth, flexibility, and low power. Further, computational platforms may replace lossy, wired interconnects with high-speed wireless interconnects. Together, data center and computational platform improvements extend the reach of cloud computing through new non-traditional wireless applications. Cellular systems may incorporate mmWave to provide higher bandwidths to solve the spectrum crunch by providing mobile networks, peer-to-peer data transfers, and backhaul in the same bands. Not all emerging applications of 60 GHz and mmWave wireless devices, however, are unprecedented. Backhaul wireless links, broadband cellular communication, intra-vehicular communication, inter-vehicular communication, and aerospace communication have all been the subject of research and some market developments. Several technology breakthroughs at mmWave, however, hope to bring these applications to larger markets with vast capabilities.
1.3.1 Data Centers
To accommodate continued growth in the Internet and cloud-based applications, Internet service providers and major Web portals are building thousands of data centers each year. Data centers are used by all major Internet companies, including Google, Microsoft, Yahoo, and Amazon, to distribute processing, memory storage, and caching throughout the global Internet. As multimedia content, for example, high-definition movies, increasingly streams over the Internet, data center buildout continues to accelerate. The buildout of data centers is comparable to the rapid buildout of towers in the early years of the cellular telephone industry.
Individual data centers often provide thousands of co-located computer servers [BAHT11]. Each data center can consume up to 30 megawatts of power, equivalent to the power drain of a small city, and must be built near a large water source (such as a lake or river) to accommodate cooling requirements. Remarkably, over 30% of the power dissipation in a typical data center is for cooling systems, for switching bottlenecks, and for broadband communication connections/circuitry between servers. Broadband circuitry is likely to become problematic as the Internet continues to expand over both wired and wireless connections [Kat09].
There are three types of communication in data centers: chip-to-chip, shelf-to-shelf, and rack-to-rack (less than 100 m). At present, data centers employ wired connections for all three types of data communication. Shelf-to-shelf and rack-to-rack communication is implemented using electrical copper connections and is the biggest bottleneck at present. Table 1.1 compares different copper solutions in terms of their power per port, reach, and link costs.
Table 1.1 A representative sample of technology choices for computer interconnections within a data center. This table makes the case for a different interconnect technology [Hor08].
Solution |
Power per Port (W) |
Port Type |
Reach |
Interconnect |
Link Cost |
CX4 |
up to 1.6W |
Dedicated copper SAS SFF8470 |
upto 15m |
4 lanes of 3.125G copper in heavy-gauge casing |
$ 250 |
10GBASE-T |
~4W |
Dedicated copper RJ45 |
30m (or 100 m) |
CAT5/CAT6 copper cable |
$ 500 |
Active Twin-ax |
1W |
Hot pluggable SFP + or XFP |
up to 30m |
Thin-gauge twin-ax copper Cable |
$ 150 |
10GBASE-SR |
1W |
up to 300m |
Hot pluggable SFP + or XFP |
Optical glass flber |
$ 500 |
Solution |
Power per Port (W) |
California Elec $/kWh |
Cost per Year |
CO2 per Year per 1600 Ports (ton) |
OPEX Cost per Year per DataCenter Cluster ($K) |
CX4 |
up to 1.6W |
20.72 |
$ 291 |
17 |
465 |
10GBASE-T |
~4W |
20.72 |
$ 727 |
42 |
1162 |
Active Twin-ax |
1W |
20.72 |
$ 182 |
11 |
291 |
10GBASE-SR |
1W |
20.72 |
$ 182 |
11 |
291 |
The broadband wired connections within data centers will not be able to accommodate future bandwidth requirements due to the increase in metal wire signal loss with increasing frequency. Data centers are expected to make a transition to other technologies. For example, Fig. 1.14 shows how optical interconnects have cost and power advantages over copper interconnects for longer ranges and/or higher power. Both cable technologies, though, have disadvantages. For example, electrical connections typically have lower bandwidth and higher dielectric losses in FR4 (a common material used for the construction of printed circuit boards), whereas optical connections typically are not standardized and installation may be costly.
)
Figure 1.14 Comparison between optical and electrical performance in terms of cost and power for short cabled interconnects. The results show that optical connections are preferred to electrical copper connections for higher data rates, assuming wires are used [adapted from [PDK+07]© IEEE].
Table 1.1 A representative sample of technology choices for computer interconnections within a data center. This table makes the case for a different interconnect technology [Hor08].
MmWave wireless communication using 60 GHz is an alternative to wired connections in data centers that could offer lower cost, lower power consumption, and greater flexibility. For example, a 10 m wireless 60 GHz link has a power budget in which 200 mW is dissipated before the power amplifier (e.g., by mixers or a voltage-controlled oscillator), 200 mW dissipated by the transmitter/antenna power amplifiers, and 600 mW of power dissipated in the channel/antennas giving a total of 1 W [Rap12b] [Rap09], which is comparable to the solutions in Table 1.1. A wireless solution allows flexible design of the data center, for example, placement of the servers, and permits easy reconfiguration. More flexible designs and a reduction in the numbers of cables and conduits allow for better placement of heat sources, and in turn, results in less stringent cooling and power requirements.
1.3.2 Replacing Wired Interconnects on Chips
The integrated antennas used to link individual 60 GHz devices may serve as the precursor of antennas used to link different components on a single chip or within a package, or within a close proximity as illustrated in Fig. 1.15. These links may be used for power combining, or more critically, signal delivery. On-chip antenna connections for power combining were evaluated as early as the mid 1980s [Reb92], but the market for high-frequency systems was limited and the technology was ahead of its time. Many researchers have experimented with on-chip or in-package wireless signal delivery (i.e., wireless interconnects) using highly integrated antennas [OKF+05]. This research demonstrates several challenges facing digital circuit design, including clock skew [FHO02] and interconnect delay [ITR09].2,3 Of these challenges, interconnect delay may be the most important to consider. The International Roadmap for Semiconductors (ITRS) identified interconnect delay as the most critical phenomenon affecting high-performance products [ITR09].
)
Figure 1.15 MmWave wireless will enable drastic changes to the form factors of today’s computing and entertainment products. Multi-Gbps data links will allow memory devices and displays to be completely tetherless. Future computer hard drives may morph into personal memory cards and may become embedded in clothing [Rap12a][Rap09][RMGJ11].
The bandwidth of copper interconnects used on-chip is also an important issue. When clock frequencies increase, the passband bandwidth is decreased due to the increased resistance exhibited by a metal wire as frequency increases. This is exhibited by the square-root dependence of metal surface resistance on frequency in addition to the skin and proximity effects that also increase resistance. An on-chip or in-package antenna may mitigate these challenges because it would reduce the total length of wire seen by a signal. Therefore, the antennas developed for 60 GHz systems may provide value across many future applications requiring very high data rates within a chip or package.
1.3.3 Information Showers
With massive mmWave spectrum and low-cost electronics now available for the first time ever, the transfer of information will become truly ubiquitous and virtually unlimited. By replacing copper wiring with massive bandwidth radio links that are located at building entrances, hallways, roadway on-ramps, and lampposts, it will soon be possible to beam entire libraries of information to people as they walk or drive. Consider today’s student, who carries a heavy backpack full of books between classes. By using a concept known as the information shower, enormous amounts of content may be transferred in seconds, with or without the student’s knowledge, as illustrated in Fig. 1.16.
)
Figure 1.16 Future users of wireless devices will greatly benefit from the pervasive availability of massive bandwidths at mmWave frequencies. Multi-Gbps data transfers will enable a lifetime of content to be downloaded on-the-fly as users walk or drive in their daily lives [Rap12a][Rap09][RMGJ11].
Memory storage and content delivery will be revolutionized using the information shower, making real-time updates and access to the latest versions of books, media, and Web content appear seamless and automatic. The student of the future will merely need a handheld communicator to obtain all of the content for her entire educational lifetime, downloaded in a matter of seconds, and updated through continual access to information showers. Furthermore, peer-to-peer networking will enable very close range wireless communications between different users, so that massive downloads by an individual user may be shared to augment content of another nearby user. Information showers will exploit both cellular and personal area networks so that future consumers of content may use low-power and lightweight devices that will replace today’s bulky and power-hungry televisions, personal computers, and printed matter.
1.3.4 The Home and Office of the Future
As mmWave devices and products evolve over the next couple of decades, the way in which our homes and offices are wired will radically change. As content from Web servers moves closer to the edge of the network, the bandwidth carried around our homes and enterprises will skyrocket by orders of magnitude. Also, the number of wireless devices that we rely upon will increase dramatically [Rap11][RMGJ11]. Today’s Internet cables will likely be replaced with massive-bandwidth mmWave radio networks, obviating the need for wired ports for Internet and telephone service, as shown in Fig. 1.17. Many low-power wireless memory devices will replace books and hard drives that are bulky and inefficient. Untethered access to information within a room and between rooms will become the norm, as humans adapt to the renaissance of wireless communications, in which our personal devices are linked by massive-bandwidth data links that carry tens of gigabits of data per second. Even today’s building wiring (e.g., Cat6 Ethernet cables) will be replaced by low-cost, high-bandwidth, rapidly deployable wireless systems that have switchable beams to adapt coverage and capacity for any building floor plan. Later chapters of this text provide the technical details needed to engineer such systems.
)
Figure 1.17 The office of the future will replace wiring and wired ports with optical-to-RF interconnections, both within a room and between rooms of a building. UWB relays and new distributed wireless memory devices will begin to replace books and computers. Hundreds of devices will be interconnected with wide-bandwidth connections through mmWave radio connections using adaptive antennas that can quickly switch their beams [Rap11] [from [RMGJ11]© IEEE].
1.3.5 Vehicular Applications
There are many applications of mmWave in the context of vehicles. Broadband communication within an automobile is being pursued to remove wired connections of vehicular devices (e.g., wires between dashboard DVD player and backseat displays) as well as to provide multimedia connectivity of portable devices inside the vehicle (e.g., MP3 players, cellphones, tablets, laptops). MmWave is especially attractive for intravehicle communications due to its inability to easily penetrate and interfere with other vehicular networks (due to high vehicle penetration losses). There are other applications outside the vehicle, as illustrated in Fig. 1.18. Vehicle-to-vehicle (V2V) communication may be used for collision avoidance or to exchange traffic information. Vehicle-to-infrastructure (V2I) links may also be used to communicate traffic information or to provide range and coverage extension of mobile broadband networks. Realization of intervehicle communication at mmWave is challenging due to the high Doppler and variable PHY and MAC conditions, which increase overhead for maintaining links, and lower transmitter height above ground, which limits the distance between automobiles of a connected network. While the current vehicle-to-vehicle standard IEEE 802.11p uses a 5.9 GHz band allocated for intelligent transportation systems, mmWave transmission is already employed at 24 and 77 GHz for automotive radar and cruise control. This makes it foreseeable that mmWave will find its way into other vehicular applications in the coming years.
)
Figure 1.18 Different applications of mmWave in vehicular applications, including radar, vehicle-to-vehicle communication, and vehicle-to-infrastructure communication.
1.3.6 Cellular and Personal Mobile Communications
Today’s cellular networks throughout the world use frequencies in the UHF and low microwave spectrum bands, between 400 MHz and 4.0 GHz. The use of these relatively low frequency spectrum bands has not changed in the 40 years of the cellular radio industry [RSM+13]. Even today, tiny slivers of spectrum (e.g., tens of MHz) within these bands continue to be allocated by governments around the world for the deployment of the fourth generation (i.e., 4G) of cellular technologies based on the LTE standard.
Demand for cellular data, however, has been growing at a staggering pace, and capacity projections are clear — cellular networks will require much greater spectrum allocations than have ever been available before. Conservative estimates of per-user data consumption growth range from 50% to 70% per year. Some wireless carriers, such as China Mobile, are already reporting even greater data consumption increases (e.g., 77% per year increase in data consumption per user from 2011 to 2012), and operators continue to experience incredible increases in video and live streaming traffic on their networks. This trend will only accelerate with time, especially as new social networking and machine-to-machine applications evolve, and as the Internet of things becomes a reality [CIS13].
The wireless community is steadily beginning to realize that the radio propagation at mmWave frequencies (dubbed “Beyond 4G,” and called “5G” by some early researchers) may not only be viable, but may actually have greater benefits than today’s cellular networks, when one considers the ability to use miniature, high-gain directionally steerable antennas, spatial multiplexing, new low-power electronics, advanced signal processing, and dormant or lightly used spectrum bands that have many tens of gigahertz of bandwidth available to them. The key technological components are about to become mature to enable multi-Gpbs mobile data rates for future mmWave wireless networks using cellular radio architectures.
Recent capacity results show that future mmWave cellular networks may use 1 or 2 GHz channels, instead of LTE’s 40 MHz RF channel bandwidths, and by using Time Division Duplexing (TDD) in a relatively small cell (200 m radius) scenario, end-user data rates will easily be increased by a factor of 20 over most LTE networks, enabling multi-Gbps mobile links for cellphone users [RRE14].
As shown in the remaining chapters of this textbook, particularly in Chapters 3-8, the frequencies above 10 GHz are a new frontier for the cellular communications field, as many orders of magnitude greater bandwidth are available for immediate use. The smaller wavelength of mmWave cellular will enable great capacity gains by exploiting spatial and temporal multipath in the channel, in a far greater manner than today’s 4G wireless networks. When additional capacity gains from beamforming and spatial multiplexing are combined with the vastly larger channel bandwidths available at mmWave carrier frequencies, it is clear that low-cost, UWB mobile communication systems with data rates and system capacities that are orders of magnitudes greater than today’s wireless networks will evolve.
Such advances in capacity are not only required as today’s cellular users demand more video and cloud-based applications, they are also logical when one considers that fact that advances in Moore’s law have brought similar order-of-magnitude increases to computer clock speeds and memory sizes over the past four decades. Wireless communications, and cellular and WiFi networks in particular, are about to realize massive increases in data rates through the use of much more bandwidth than ever available before, and with this massive bandwidth will come new architectures, capabilities, and use cases for cellphone subscribers [PK11]. Such advances will usher in the renaissance of wireless communications [Rap12a].
1.3.7 Aerospace Applications
Because of the significant absorption of signals in oxygen, the 60 GHz spectrum is ideal for aerospace communication where terrestrial eavesdropping must be avoided [Sno02].4 Consequently, many spectrum regulations, including FCC regulations in the USA [ML87], have allocated 60 GHz for intersatellite communication. Intersatellite communication links are LOS, and special design considerations for satellite systems result in few technology translations to consumer applications. One emerging 60 GHz aerospace application is multimedia distribution in aircraft, as illustrated in Fig. 1.19, to reduce the cabin wiring [GKT+09]. The localization of 60 GHz signals and the massive bandwidth resources make 60 GHz attractive versus microwave frequencies [BHVF08]. Unfortunately, to protect intersatellite communication from wireless in-aircraft applications, regulations currently disallow 60 GHz wireless communication in aircraft. Regulations, however, are likely to change in the future with enough industry pressure and demonstration of the feasibility of network coexistence. Also, as mmWave wireless becomes more mature, additional high-attenuation bands, such as 183 and 380 GHz, will find use in aerospace applications.
)
Figure 1.19 Different applications of mmWave in aircraft including providing wireless connections for seat-back entertainment systems and for wireless cellular and local area networking. Smart repeaters and access points will enable backhaul, coverage, and selective traffic control.