C.10 Portability, Mobility and Ubiquity

Principal Authors:

Brian Bershad and John McDonald

Additional Contributors:

Joel Birnbaum, David Borth, Daniel Duchamp, Larry Dunn, William G. Griffin, Bill Hamer, Jay Kistler, Yuet Lee, Brian Marsh, Bruce Mendel, John A. Morse, Dave Paldan, Tim Regan, Daniel Siewiorek, Tim Stone, Robert Vence, Werner F. Wedam, Stephen Weinstein, Mark Weiser and John Wohlstetter

1. Introduction

The National Information Infrastructure will make it possible to provide portable, mobile and ubiquitous access to services. By portable, we mean that people may easily move from site to site yet still have access to a rich set of services wherever they are. By mobile, we mean that users will be able to access services while on the move, taking advantage of a wireless networking infrastructure. By ubiquitous, we mean that access to computing, communication and information services will be as common as today's telephone and that the availability of a nationwide networking infrastructure will make it possible to bring connectivity to a wide range of electronic devices.

This section discusses the technical demands that will be imposed on the NII by the requirements of PMU (portability, mobility and ubiquity) services. It is structured in three parts. In Section 2, we discuss four PMU application areas in order to illustrate the kinds of services that people will use. Second, we use those areas to motivate a list of technical requirements that should be satisfied by NII research. Third, we recommend a set of possible research programs from which the requirements would be satisfied.

2. Motivation for Portability, Mobility and Ubiquity

Four application domains that will benefit from the PMU component of the NII are:

Portable office: Enable people to function outside of the office.
Crisis management: Deal with unexpected emergencies in the field.
Personal services: Inform, entertain and interconnect the population.
Distributed measurement/sensor systems: Collect and manage large amounts of remote information.

These domains relate to the five major service areas of the NII: education, manufacturing, health care, finance and government.

Other application domains certainly exist.

2.1 Application Domain: Portable Office

The portable office will allow workers and students to coordinate their activities from anywhere. Today, the portable office represents both the least and most successful of the PMU areas discussed in this section. It is the most successful in that large numbers of people are currently engaged in some sort of "portable office" computing. By 1996, portable computers will account for over half of all computers sold in the United States. On the other hand, the frustration faced by those who are trying to be truly portable points to the lack of infrastructure that supports the portable office.

Several non-obvious application areas enabled by an NII-based PMU infrastructure include:

Education

Students and teachers maintain offices in a strict, although unconventional sense. They travel between home and school, and work in both places. A PMU infrastructure will allow students and teachers to do their work regardless of location.

Manufacturing

The manufacturing process has many stages including product planning, design, forecasting, production, delivery and service. Productivity in each of these steps would be enhanced if workers could avoid the cost and time relating to commuting for even a small portion of their time.

Financial Trading

The financial trader may participate in transactions from anywhere, whether on the floor of Wall Street or while traveling from customer site to customer site. The trader requires not only market-independent information, such as client names and account balances, but also market-based information, such as price quotes.

2.2 Application Domain: Crisis Management

The NII will enable crisis situations, such as natural disasters, to be effectively managed. A crisis management team must be able to rapidly install high-performance equipment, must be able to rapidly access information for geographically dispersed databases, and must be able to prioritize and analyze the information and available resources in order to address the crisis at hand.

Several application areas in which crisis measurement services can be used to improve the quality or performance of the crisis management team include:

The Mobile Clinic

In emergencies such as train wrecks, earthquakes or industrial accidents there must be a mobile emergency room with remote access to a global database including patient history. Remote consultation may be needed.

The Firefighter

Fire Emergency response can be improved through the delivery of information. Firefighters require information on building location, the specific fire location, fire type, water or chemical access points, building access points, team member location, hazardous material storage conditions in the vicinity, civilian personnel in the vicinity and more. This information needs to be timely and relevant to the emergency situation. This information must be prioritized, appropriately filtered and delivered to the response team at the site and during transit to the site.

2.3 Application Domain: Personal Information Services

The NII will make possible a broad range of personal information services. Currently, personal information services for most people are exemplified by television, telephone and the post office. The NII will provide broad-scale, but more efficient and/or entertaining, versions of these services.

The applications in this domain are grouped under two major classes: interpersonal communications and personalized information delivery.

Interpersonal communication involves mostly peer-to-peer interaction, and can be either synchronous or asynchronous. Synchronous communication includes telephony and videotelephony, application sharing, computer conferencing, games, gambling and other collaborative activity. Asynchronous communication includes video and voice messaging, and integrated, wide-scale electronic mail.

Personalized information delivery services will include pagers, text-on-demand, cellular phones, video-on-demand and home shopping. Bandwidth requirements will range from low to high and may be asymmetric.

2.4 Application Domain: Distributed Measurement/Sensor Systems

The area of distributed measurement creates a special challenge for PMU applications in the NII. Distributed measurement involves the collection, analysis and dissemination of data across a possibly wide geographic area. The purpose of distributed measurement is to allow small, possibly portable, possibly mobile, sensor devices to be organized as part of the fabric of everyday life. In so doing, an infrastructure that supports distributed measurement can make it possible to deploy sensor devices that collect data about the environment and transmit the data back to intelligent processing sites. At those sites, the data can be processed, stored and acted upon either by individuals or automatic programs. With this, important data can be collected at lower cost, with higher integrity, with greater longevity, and ultimately, greater value to the individuals and systems for which the data is relevant.

Several application areas in which distributed measurement can be used to improve the quality or efficiency of people's lives include:

Health Care

Distributed patient monitoring systems will enable doctors to keep track of their patients without requiring costly office visits or long-term "watching" hospitalization. Constant recording of the vital signs enables the provider to carefully monitor the progress of a recovering patient.

Environmental Sensing and Measurement

A coherently deployed distributed sensor system would allow important information such as temperature, smog levels, light levels, traffic flow, seismic activity and water quality to be inexpensively collected and processed. Currently, environmental data collection is ad hoc, ranging from agents in the field to secondary sources. By being able to collect information about the environment through remote sensor units, trouble spots can be easily identified, remotely diagnosed and repaired.

Manufacturing and Product Delivery

The manufacturing, design and delivery process could be improved through a distributed sensor and measurement network. Individual parts could be identified on the factory floor, making inventory control a more tractable problem. Just-in-time inventory control and manufacturing could be achieved through centralized management and monitoring sites. Vendors could easily sell one another goods based on immediate availability. An accessible network of inventory information would make it possible to create a trading system that spans vertical boundaries. Complex manufacturing, design and delivery processes could be augmented with sensors as needed to detect problems in the line. For example, consider the problem of diagnosing a wiring problem (for instance, in a large communications network). A ubiquitous sensor network could help pinpoint the problem as it occurs. Floor configurations can be rapidly changed without rewiring the sensor network.

3. Technical Challenges

The key challenge in providing PMU is to do so in a general way so that people and devices can exploit the NII wherever, whenever and whatever they are. Today, we can identify specialized instances of PMU, such as cellular phones, automatic call forwarding and laptop computers, but there does not exist any underlying fabric with which these instances can be tied together or leveraged. The absence of this infrastructure prevents all but the heartiest (or the wealthiest) from exploiting the NII in a location-independent manner.

We have broken the application requirements into the following seven categories: networking, session and application layer, security, location management, accounting, user interfaces, and resource and device management.

3.1 Networking

Inexpensive wired and wireless network connections.
Support for a wide range of bandwidths.
Universal connectivity that will enable equipment to interface with differing communications equipment, protocols, bandwidth, frequency and capability.
Smooth internetworking between the wired and wireless worlds.
Video source coding matched to the impairments of wireless channels, and highly compressive coding schemes.
Real-time guarantees for continuous media streams in wireless networks. Many applications, for example those in the medical area, require bounded delay when delivering a packet to the health care provider.
Quality-of-service guarantees in terms of bandwidth, latency and error/dropout rate are required for different media types, channel characteristics and traffic conditions.

3.2 Session and Application Layer

Rapid and context-relevant access to globally dispersed databases organized and prioritized by context.
Operation under conditions of partial, intermittent or degraded connectivity, together with intermediate transitions.
A universal time stamp mechanism to coordinate input data from multiple sources.
A standard protocol suite to ensure interoperability.
Effective file synchronization techniques to enable disconnected operation.
A standard technology for downloading new protocols and programs into devices.

3.3 Security

Secure authentication and authorization mechanisms because few guarantees about physical integrity can be provided.
Users should be able to hide their presence and activity from all, from a designated set of people and, in some cases, from no one, as they require.

3.4 Location Management

Fine-grained location sensing and processing to facilitate location-dependent applications.

3.5 Accounting

A flexible fee structure to enable a large number of devices with both low and high bandwidth requirements.

3.6 User Interfaces

A UI for PMU should have low resource requirements, but should also support a range of input modes such as hands-free operation (requiring voice and gesture input).

3.7 Resource Management and Device Architecture

Special application requirements relating to PMU apply to terminals. Component technology must fit the application, e.g. batteries and displays.

4. Research and Development Recommendations

Below we recommend several research programs for which the goals are to produce results that satisfy the above requirements. Some of the above requirements will be addressed by other sections of this report; for example, user interface technology is addressed in the Ease of Use section.

Power and configuration management to explore optimizations of coding/power/bandwidth/cell size trade-offs in microcells.
Portable terminal componentry with special emphasis on batteries, displays and I/O.
Wireless technology to develop machinery for integrated two-way wireless systems, including research in both component and systems aspects of cheap and quick-connecting yet dependable uplinks. Nodes may be remote, portable or fixed. This research would also develop frequency-adaptive radio protocols.
Protocols for a wide range of asymmetric networks to ensure that the NII is not overly biased toward applications with symmetric, or excessively high or excessively low bandwidth requirements.
Protocols for mobile networks to define network and transport level interfaces that function across diverse physical networks, that offer real-time guarantees and that are appropriate for both low-end devices and high-end mobile/portable computers. Network protocols should be robust to handle higher rates and dropouts due to hand-offs or disconnects.
Compression technology to design algorithms and protocols that are more robust for wireless transmission to deal with higher error rates, smaller packet lengths, etc.
Location management technology would develop services for inexpensive location sensing and processing, especially indoors.
Security and time protocols would develop and deploy mechanisms for digital signature, encryption and universal time stamps.
Identification, authorization and authentication mechanisms by non-invasive means. Fingerprint and retina examination technology exists, but these techniques may not be suitable nor available. Research into alternative universal methods should be undertaken.
Replicated data management to develop general caching and consistency mechanisms for distributed data and disconnected operation.
Programming paradigms for weakly connected environments to develop general techniques for writing programs that work in environments with variable connectivity.
Database access for a range of data types, including multimedia, and access and filing methods, including relational and object-oriented.
Small size, low cost, non-volatile storage with fast access, high data rates and low power requirements for geographically dispersed and mobile databases.