EMNTG also develops and customizes protocol-level models of wireless technologies such as LTE. These models can be used to perform more detailed simulations of an incident area communication scenario. They include models of the application traffic (e.g., push-to-talk voice, video streams, file transfers), lower layer communication protocols, and the RF channel environment. Performance metrics generated by these models include
• Achieved throughput
• Communication delay
• Packet loss rate
• Network utilization

Protocol models complement the wide area network analysis and optimization tools described above to provide a more complete picture of the expected public safety communication performance. Using the cell sizes, antenna configurations, and transmission powers obtained from a network optimization, the protocol-level analysis generates realistic sector loads which can be fed back to the network analysis and optimization tools for further refinement of the optimized configuration.

Channel Measurements
The fidelity of the predictions generated by the network and protocol models is dependent in large part on the accuracy of the underlying RF channel propagation model. In collaboration with its partners in NIST/PML, EMNTG collects and analyzes channel measurements that are used to develop and tune RF propagation models. Of particular relevance to the Public Safety Broadband Network, these efforts include measurements in the 700 MHz public safety band.
Multiservice broadband networks are on the verge of becoming a reality. The key drivers are rapid advances in technologies; privatization and competition; the convergence of the telecommunication, data communication, entertainment, and publishing industries; and changes in consumers' life-styles.
Wide-area networks today consist of separate networks for voice, private lines, and data services, supported by a common facility infrastructure. Similar separation between voice and data networks exists on enterprise premises (Fig. 1). Residential users typically use copper loop to access voice networks as well as data networks (using voice-band modems).



Public voice and private line networks have been designed for very low latency and with an unrelenting attention to reliability and quality of service. Data networks (especially the public Internet) introduce longer and less predictable delays and are not suitable for highly interactive communication. For the most part, the Internet has not yet been shown to be reliable enough for mission-critical functions.
The time is ripe for these networks to change in fundamental ways. There is a demand for services involving multiple media, increasing intelligence, diverse speeds, and varied quality-of-service requirements. Increasing dependence on network services requires all forms of networking to be as reliable as today's voice network. The revolution taking place in electronics and photonics will provide ample opportunities to satisfy these requirements.
Technological advances
The storage capacity of a single dynamic read-only-memory (DRAM) chip increased from 64 kilobits in 1970 to 256 megabits in 1998, and 4-gigabit chips are possible in research laboratories (Fig. 2 a). Development of multistate transistors and new lithographic techniques (enabling the fabrication of atomic-scale transistors) promise the continuation of this trend. The computing power in a single microprocessor chip increased from 1 million instructions per second (MIPS) in 1981 to about 400 MIPS in 1998, and no limit is in sight (Fig. 2 b). These advances are being translated into explosive growth in switching and routing capacities, information processing capacities, database sizes, and data retrieval speeds. Atomic-scale transistors also promise “system on a chip,” resulting in inexpensive wearable computers, smart appliances, and wireless devices. They also promise less power consumption and longer battery life.

Fig. 2 Progress in semiconductor technologies. (a) Density of dynamic random-access memories (DRAMs). Lengths (in micrometers) are minimum internal spacings of chip components. Advances in manufacturing techniques allow narrower spacing, making
possible higher density and hence larger storage capacity. (b) Microprocessor speeds.
The advances in photonics are even more remarkable. The capacity of a single optical fiber increased from 45 megabits per second in 1981 to 1.7 gigabits per second in 1990 (Fig. 3). A major change occurred with the advent of dense wavelength-division multiplexing (DWDM) and optical amplifiers. Dense wavelength-division multiplexing allows the transport of many colors (wavelengths) of light in one fiber, while optical amplifiers amplify all wavelengths in a fiber simultaneously. These two innovations have made it possible to carry 400 gigabits per second (40 wavelengths each carrying 10 gigabits per second) on a single fiber and 1600 Gbps systems are on the horizon. Experimental work in research laboratories is pushing this capacity to over 3 terabits per second. Recent innovations have made it possible to put 432 fibers in a single cable. One such cable will be able to transport the daily volume of current worldwide traffic in 60 seconds. Many new and existing operators of wide-area networks have already started capitalizing on this growth in transport capacity. Transport capacity in wide-area networks around the world is expected to increase by a factor of up to 1000 by 2005.

Fig. 3 Capacity growth in long-haul optical fibers. Evolution is plotted in both the total capacity and the number of wavelengths employed.
Advances in electronics and photonics will also permit tremendous increases in access speeds, releasing the residential and small business users from the current speed restriction of the “last mile” on the copper loop that links the user to the network. New digital subscriber loop (DSL) technologies use advanced coding and modulation techniques to carry from several hundred kilobits per second to 50 megabits per second on the copper loop. Hybrid fiber-coaxial (HFC) technologies allow the use of cable television channels to provide several megabits per second upstream and up to 40 megabits per second downstream for a group of about 500 homes. Similar access speeds may also be provided by hybrid fiber-wireless technologies. Fiber-to-the-curb (FTTC) and fiber-to-the-home (FTTH) technologies provide up to gigabits per second to single homes by bringing fiber closer to the end users. Passive optical networks (PONs) will allow a single fiber from the central hub to serve several hundred homes by providing one or more wavelengths to each home using wavelength splitting techniques.
Advances in network infrastructure
Innovations in software technologies, network architectures, and protocols will harness this increase in networking capacities to realize true multiservice networks.
Narrowband voice [analog, ISDN (Integrated Services Digital Network), and cellular] and Internet Protocol (IP)–based data end systems are likely to continue operating well into the future, as are the Ethernet-based local-area networks. Most of the present growth is in IP-based end systems and cellular telephones, although many parts of the world are still experiencing enormous growth in analog and ISDN telephones. New services are being developed for the IP-based end systems (such as multimedia collaboration, multimedia call centers, distance learning, networked home appliances, directory assisted networking, on-line language translation, and telemedicine). These services will coexist with the traditional voice and data services. See also: Distance education

One challenge is to find the right network architecture for transporting efficiently the traffic generated by current and new services. An approach involving selective layered bandwidth management (SLBM) provides efficiency, robustness, and evolvability. SLBM uses all or some of the IP, Asynchronous Transfer Mode (ATM), SONET/SDH (synchronized optical network/synchronized digital hierarchy), and optical networking layers, depending on the traffic mix and volume (Fig. 4).


Fig. 4 Protocol layering and evolution. Heavy arrows represent new and future layering. PPP is Point-to-Point protocol; HDLC, High-speed Data-Link Control (used here only for


Fig. 4 Protocol layering and evolution. Heavy arrows represent new and future layering. PPP is Point-to-Point protocol; HDLC, High-speed Data-Link Control (used here only for the purpose of delineating the packets); SDL, Simplified Data-Link protocol; HSSF, High-Speed Synchronous Framing protocol; Cell PL, (ATM) Cell-based Physical Layer.





For the near future, ATM provides the best technology for a true multiservice backbone due to its extensive quality-of-service and traffic management features such as admission controls, traffic policing, service scheduling, and buffer management. An ATM backbone network can support the traditional voice traffic, new packet voice traffic, various types of video traffic, Eithernet traffic, and IP traffic. The ATM Adaptation Layer protocols AAL1, AAL2, and AAL5 facilitate this integration. ATM also enables creation of link-layer (layer 2) virtual private networks (VPNs) over a shared public network infrastructure while providing secure communication and performance guarantees.
Traditional voice, IP, and ATM traffic will be transported over the circuits provided by the SONET/SDH network layer over the optical (wavelength) layer. SONET/SDH networks allow partitioning of the wavelength capacity into lower-granularity circuits which can be used by this traffic for efficient networking. The SONET/SDH layer also provides extensive performance monitoring. Finally, SONET/SDH networks, formed using ring topology, allow simple and fast (several tens of milliseconds) rerouting around a failure in a link or a node. This fast restoration of service makes such failures transparent to the service layers. The optical layer will consist of point-to-point fiber links over which many wavelengths will be multiplexed to allow very efficient use of the fiber capacity. Many public carriers and enterprise networks will use this multilayered networking. However, all these situations will change over time.
Rapid growth of IP end systems and high access speeds to the wide-area networks will create high point-to-point IP demands in the core backbone. At this level of traffic, the finer bandwidth partitioning provided by the ATM layer may not compensate for the additional protocol and equipment overhead. Thus, an increasingly higher portion of IP traffic will be carried directly over SONET/SDH circuits. At even higher traffic levels, even the partitioning provided by the SONET/SDH layer will be unnecessary. Meanwhile, optical cross connects and add-drop multiplexers will allow true optical-layer networking using ring or mesh topology. New techniques and algorithms being developed for very fast restoration at the optical layer will further reduce the need for the SONET/SDH layer. IP-over-wavelength networking is then expected to develop.
Of course, this simplification in the network infrastructure requires IP- and Ethernet-based networks to multiplex voice, data, video, and multimedia services directly over the SONET/SDH or dense wavelength-division multiplexing layer. Ethernet and IP protocols are being enhanced to allow such multiplexing.
New high-capacity local-area network switches (layer 2/3 switches) provide fast (10–1000 megabits per second) interfaces and large switching capacities. They also eliminate the need for contention-based access over shared-media local-area networks. Protocol standards (802.1p and 802.1q) are being defined to provide multiple classes of service over such switches by providing different delay and loss controls. Similarly, router technology and Internet protocols are beginning to eliminate the bottlenecks caused by software-based forwarding, rigid routing, and undifferentiated packet processing. New IP switches (layer 3 switches) use hardware-based packet forwarding in each input port, permitting very high capacities (60–1000 gigabits per second, or 50–1000 million packets per second). Protocols are being defined to use the Differential Service (DS) field in the header of each IP packet to signal the quality-of-service requirement of the application being supported by that packet. Intelligent buffering and packet scheduling algorithms in IP switches can use this field to provide differential quality of service as measured by delay, jitter, and losses. At the interface between layer 2 and layer 3 switches, conversion between 802.1p- and DS-based signaling can provide seamless quality-of-service management within enterprise and carrier networks. Resource reservation protocols are being defined to allow the application of signal resource requirements to the network. Traffic-policing algorithms will guarantee that the traffic entering the network is consistent with the resources reserved by the reservation protocols. Resource reservation protocols and traffic policing will further help quality-of-service management. Two other factors will add to the quality-of-service capability of IP networks. In particular, hierarchical classification of IP traffic using additional fields in the header will allow bandwidth guarantees and delay control for individual or aggregated flows based on the users as well as applications. Also, ongoing work on Multi Protocol Label Switching (MPLS) will allow flexible routing and traffic engineering in IP networks.
With these capabilities, public carriers can offer IP-layer (layer 3) virtual private networks to many enterprise customers over a shared infrastructure. Carrier-based policy servers, service and resource provisioning servers, and directories will interact with enterprise policy servers to map the enterprise requirements into carrier actions consistent with the overall virtual private network contract.
While backbone networks are likely to become more uniform, many different access technologies (such as copper loop, wireless, fiber, coaxial cable, fiber-cable hybrid, and satellite) will continue to play major roles in future networks. Standards are being defined for each of the access technologies to support diverse quality-of-service requirements over one interface. Thus, a true end-to-end solution is expected for multiservice broadband networking.
See also: Data communications; Integrated circuits; Integrated services digital network (ISDN); Local-area networks; Microprocessor; Optical communications; Packet switching; Semiconductor memories; Telephone service; Wide-area networks

New technology uses human body for broadband networking
Your body could soon be the backbone of a broadband personal data network linking your mobile phone or MP3 player to a cordless headset, your digital camera to a PC or printer, and all the gadgets you carry around to each other.
These personal area networks are already possible using radio-based technologies, such as Wi-Fi or Bluetooth, or just plain old cables to connect devices. But NTT, the Japanese communications company, has developed a technology called RedTacton, which it claims can send data over the surface of the skin at speeds of up to 2Mbps -- equivalent to a fast broadband data connection.
Using RedTacton-enabled devices, music from an MP3 player in your pocket would pass through your clothing and shoot over your body to headphones in your ears. Instead of fiddling around with a cable to connect your digital camera to your computer, you could transfer pictures just by touching the PC while the camera is around your neck. And since data can pass from one body to another, you could also exchange electronic business cards by shaking hands, trade music files by dancing cheek to cheek, or swap phone numbers just by kissing.
NTT is not the first company to use the human body as a conduit for data: IBM pioneered the field in 1996 with a system that could transfer small amounts of data at very low speeds, and last June, Microsoft was granted a patent for "a method and apparatus for transmitting power and data using the human body."
But RedTacton is arguably the first practical system because, unlike IBM's or Microsoft's, it doesn't need transmitters to be in direct contact with the skin -- they can be built into gadgets, carried in pockets or bags, and will work within about 20cm of your body. RedTacton doesn't introduce an electric current into the body -- instead, it makes use of the minute electric field that occurs naturally on the surface of every human body. A transmitter attached to a device, such as an MP3 player, uses this field to send data by modulating the field minutely in the same way that a radio carrier wave is modulated to carry information.
Receiving data is more complicated because the strength of the electric field involved is so low. RedTacton gets around this using a technique called electric field photonics: A laser is passed though an electro-optic crystal, which deflects light differently according to the strength of the field across it. These deflections are measured and converted back into electrical signals to retrieve the transmitted data.
An obvious question, however, is why anyone would bother networking though their body when proven radio-based personal area networking technologies, such as Bluetooth, already exist? Tom Zimmerman, the inventor of the original IBM system, says body-based networking is more secure than broadcast systems, such as Bluetooth, which have a range of about 10m.
"With Bluetooth, it is difficult to rein in the signal and restrict it to the device you are trying to connect to," says Zimmerman. "You usually want to communicate with one particular thing, but in a busy place there could be hundreds of Bluetooth devices within range."
As human beings are ineffective aerials, it is very hard to pick up stray electronic signals radiating from the body, he says. "This is good for security because even if you encrypt data it is still possible that it could be decoded, but if you can't pick it up it can't be cracked."
Zimmerman also believes that, unlike infrared or Bluetooth phones and PDAs, which enable people to "beam" electronic business cards across a room without ever formally meeting, body-based networking allows for more natural interchanges of information between humans.
"If you are very close or touching someone, you are either in a busy subway train, or you are being intimate with them, or you want to communicate," he says. "I think it is good to be close to someone when you are exchanging information."

RedTacton transceivers can be treated as standard network devices, so software running over Ethernet or other TCP/IP protocol-based networks will run unmodified.
Gordon Bell, a senior researcher at Microsoft's Bay Area Research Center in San Francisco, says that while Bluetooth or other radio technologies may be perfectly suitable to link gadgets for many personal area networking purposes, there are certain applications for which RedTacton technology would be ideal.
"I recently acquired my own in-body device -- a pacemaker -- but it takes a special radio frequency connector to interface to it. As more and more implants go into bodies, the need for a good Internet Protocol connection increases," he says.
In the near future, the most important application for body-based networking may well be for communications within, rather than on the surface of, or outside, the body.
An intriguing possibility is that the technology will be used as a sort of secondary nervous system to link large numbers of tiny implanted components placed beneath the skin to create powerful onboard -- or in-body -- computers.