See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/228930531

Requirements for an ubiquitous computing

infrastructure

Article January 2001

CITATIONS READS

4 189

1 author:

Sergio Maffioletti
University of Zurich
43 PUBLICATIONS 192 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Emmental View project

All content following this page was uploaded by Sergio Maffioletti on 01 October 2017.

The user has requested enhancement of the downloaded file.

 Requirements for an Ubiquitous Computing Infrastructure
Sergio Maffioletti 
Department of Informatics
University of Fribourg Chemin du Musee 3
1700 Fribourg, Switzerland

Abstract
Ubiquitous computing is an emerging paradigm for interactions between people and computers. Its
aim is to break away from the desktop computing to provide computational services to a user when and
where required. In this paper we present the design criteria for an infrastructure platform for ubiquitous
computing applications. The motivation for building such infrastructure is to formalize a common design
methodology for developing application based on the notion of Interactive Environment.

1 Introduction
The way people used to think about computers and, as consequence, the way they interact with them, is
always determined by the constrains the technology impose. The more the technology availability and
reliability can be taken for granted, the more the users change the nature of their requirements. In the
early days of computation, where the technology support was rather rough and unreliable, users expecta-
tions were technology-driven. More performances and more reliable hardware was the customer s entire
requirement. Nowadays technology is entering in its mature phase[4]. This means we can start taking
technology for granted, and customers tend to change their expectations focusing on their everyday re-
quirements, like the freedom to move around, the freedom to chose the medium used to attend the services
offered by this technology. The continuous changing of the relation between the technology and the users
alter the role of technology in our lives. In the past fifty years of computation there have been two great
trends in this relationship: the mainframe and the PC relationship. Today Internet is carrying us through an
era of widespread distributed computing toward the relationship of Ubiquitous Computing [8] character-
ized by deeply imbedding computation in the world. The vision of ubiquitous computing, first expressed
by Weiser[7], is to break away from the desktop computing to provide computational services to a user
when and where required. The ubiquitous computing era will have lots of computers sharing each of us.
Some of these computers will be embedded in walls, chairs, clothing, light switches, cars. They will op-
erate in the background of our life, in a transparent and not intrusive way. This vision will change not
only the way well interface with computers but, also, the class of applications deployed for this scenario.
Actually applications deeply rely on computers structure: desktop applications are conceived to be ex-
ecuted by a PC with a standard input system like keyboard or mouse and a rather simple output system
like a monitor; distributed applications rely on a presence of more computers networked together. None of
them have users as a central notion of their model. Developing applications for UbiComp scenario means
to deal with completely different constrains and functionalities determined by the user expectations, like
the availability of services wherever, whenever and in any for the user require. This is a rather important
shift in HCI domain: applications integrated in our reality and realized by the interaction between one or
more computing devices. Well be able to use devices of different forms and functionality; well interact
with them in a more natural way (speech, vision, movement, facial expressions. . . ). The motivation for
building these systems is to bring computation into the real, physical world to support what is tradition-
ally considered non-computational activity; to allow computer to participate in activities that have never
 [email protected]

1
previously involved computation and to allow people to interact with computational system the way they
interact with other people: via gesture, voice, movement, and context. The main challenge for the success
of these systems is the design of smart user interfaces and software that allows for ubiquitous and easy
access to personal information and that is flexible enough to handle changes in user context and availability
of resources; in one word an infrastructure. In this paper we present the basic requirements for an UbiComp
infrastructure focused on defining new models for the development of interactive environment applications.

2 Ubiquitous Computing System

Ubiquitous Computing (UbiComp) denoted the universal availability of computation throughout multiple
systems in the users environment. A ubiquitous computing system consists of a heterogeneous set of
computing devices; a set of supported tasks; and some infrastructures the devices may rely on in order to
carry out their tasks. Differently form desktop systems, the emphasis is on combining software components
to provide services to the user. With an UbiComp system we are concerned not only with software services
but also with devices and how to combine them. Analyzing an UbiComp system were interested in classify
the main functionalities it supplies differently from the standard desktop systems. Weisers classification[7]
of an UbiComp system is based on two fundamental attributes:

 Ubiquity: interaction with the system is available wherever the user needs it.
 Transparency: the system is non-intrusive and is integrated into the everyday environment.

According with this classification Abowd[5] identified two dimensions that provide a rather clear
boundary for UbiComp systems and express the relationship with other emerging research areas:

 User mobility: reflect the freedom the user has to move about when interacting with the system.
 Interface transparency: applies to the systems interface and reflects the conscious effort end attention
the system requires of the user, either for operating it or for perceiving its output.

User mobility

UbiComp
mobile
some

computing

wereable
computing smart
environment
augmented
reality
none

desktop adaptive
computing interface

none some Interface transparency

Figure 1: an ontological framework

Figure 1 present an ontological framework based on the two dimensions defined above. This framework
allows characterizing other research streams relevant to HCI and ubiquitous computing. As we may see
UbiComp system tries to maximize the user mobility and the interface transparency. In contrast desktop
computing offer no user mobility and, for the most part, no transparency to the end user.
3 Interactive Environment
Research in ubiquitous computing is toward the development of application environment able to deal with
the mobility of both users and computing devices. The vision of ubiquitous computing relies on the pres-
ence of environments enriched by computers embedded in everyday objects (blackboards, table, chairs. . . )
and by sensors able to catch information form the context. It is then an important requirement for Ubi-
Comp applications to provide a support environment in which specialized computing instruments1 can be
accommodated and integrated into existing application contexts. Our interest is in Interactive Environ-
ments, defined as Intelligent Environments[1] or Cooperatives Buildings[6], conceived primary for
workgroup interaction. These environments represent the spatial boundaries of applications integrated in
our everyday context; they represent the physical space where the applications are placed and executed.
Applications developed for these environments have two dimensions: a service dimension that represents
the number of services available in the system and a device dimension that represents the number of de-
vices incorporated in the environment. Interactive environments have specific computational properties that
generally distinguish them from other computational systems. They have a large number of hardware and
software components that need to cooperate; they tend to be highly dynamic and require reconfiguration
and resource management on the fly as their components and inhabitants change and as they adjust their
operation to suit the learned presence of their user. Even though each interactive environment is created
in its own way for its own purpose, they are generally built out of similar components. Intelligent Envi-
ronments defined in the Metaglue[1] project are an example. Intelligent Environments contain a multitude
of subsystems comprising their perceptual interfaces, software applications, hardware device connections,
and mechanisms for internal control.

Application Context

Federation

Device

Service

Sensor

Figure 2: An Interactive Environment.

Figure 2 shows an example of an interactive environment. The application context represent the log-
ical area where the application takes place and corresponds to one of the physical regions composing the
interactive environment (these regions may vary in dimension from a single room to an entire building). A
federation of computing instruments is a set of authenticated instruments cooperating with each other. The
federation represents the group of instruments belonging each time to a specific application context. The
application will use the federations topology in order to determine the service-device association. Hence
an application is realized by the cooperation between services associated to different devices belonging to
the same application context.

4 Common Middleware
The few existing integrated multi-device computer environments today tend to be highly specialized and
based on application-specific software. Applications developed for interactive environments should be able
1 With computing instrument we refer to both devices and sensors. It identifies the abstract idea of computers integrated in our

everyday environment
to interconnect and manage large numbers of disparate hardware and software components. They should
operate in real-time; dynamically add and remove components to a running system without interrupting its
operation; control allocation of resources; and provide a means to capture persistent state information. Fre-
quently these components are not designed to cooperate, so not only they must be connected, but also there
is a need to express the logic of this interconnectio n. In other words, inter -component connections are
not merely protocols, but also contain the explicit knowledge of how to use these protocols. Thus, viewing
the connections simply as an application programming interface is insufficient. Cooperation among dif-
ferent applications is also difficult to achieve without a common platform. In order to model applications
in this domain we need to define a common design methodology based on new paradigms independents
from the technology. We are investigating a model to abstract the main components of an UbiComp system
in order to formalize the development of interactive environment applications. These components may be
classified into three abstraction layers:

 Physical deals with technological constrains.

 Middleware defines structure and the cooperation of abstract services.
 Application concerns the user interfaces.

Thank to these abstractions a middleware will present a uniform access abstraction for different ubiqui-
tous devices, allowing them to interact and cooperate. This allows us to write applications scaling both on
services offered and on devices composing the system. The model is not intended to be used for a single
application, but to provide a standardized view of basic interactive environment functionality.

4.1 Functionalities
The main requirements that arise from the idea of a common middleware are:

 Mobility: First of all, the middleware has to deal with the mobility of the users and other physical
objects. Therefore it must support mobile wireless communication and small mobile computing
devices, such as PDAs and laptops.
 Heterogeneity: Like in the Nexus[2] project, a common middleware has to accommodate a large
variety of application environments, ranging from areas covering whole nation to areas just covering
an office. This leads to a very heterogeneous system, which has to support different network tech-
nologies and different tracking systems for example. Further aspects of heterogeneity come from
the different services that rely on the presence of this middleware and the various input and output
devices that have to be supported.
 Scalability: The middleware has to scale well for both a large number of cooperating services, which
realize the application in each application context, and a large number of devices involved each time
the application is used. Services represent the logical dimension of the application, while devices
represent its physical dimension.

Existing projects like Oxygen[3], Nexus[2], Beach[6] and Metaglue[1] address the functionalit ies
stated above as well but do not consider higher level service classification. Applications in these projects
cannot rely on a suitable abstraction layer describing their functionalities. What we need, instead, is to
allow applications to define their own ontology for classifying resources they may offer and require. These
high level concepts will be associated to low -level service implementation by the base infrastructure ac-
cording with the constrains defined in each application context. This allows both the application and the
instruments to use a high-level service description for interaction. In such a way each computing instrument
is also able to roam from one application context to another (even if the other is using another ontology)
without changing its service description.
 5 Conclusion
In this paper we have presented the general idea of a middleware that provides the basic functionalitys for
modeling interactive environment applications. This middleware will allow such applications to be created
much more easily as they can rely on a common infrastructure. The platform maintains specific models
for certain areas of the real world, which allow a user to access information, or services by their spatial
position or through real world objects. A middleware for UbiComp applications aim at answering the need
to provide a unified vision of the different computing functionalitys issuing from convergence between
information and communication technologies.

References
[1] Michael H. Coen, Brenton A. Philips, Nimrod Warshawsky, Luke Weisman, Stephen Peters, and Peter
Finin. Meeting the Computational Needs for Intelligent Environment: The Metaglue System. In
Submitted to MANSE9, 1999.
[2] Hohl Fritz, Uwe Kubach, Alexander Leonhardi, and Kurt Rothermel. Next Century Challenges: Nexus
An Open Global Infrastructure for Spacial Aware Applications. ACM, 1999.
[3] MIT project Oxygen, June 2000. www.oxygen.les.mit.edu.
[4] Donald A. Norman. The Invisible Computer. The MIT Press, Cambridge, Massachusetts 0214, 1999.
[5] D. Salber, A. K. Dey, and G. D. Abowd. Defining an HCI Research Agenda for an Emerging Interaction
Paradigm. Technical report, GVU Center and College of Computin, Georgia Tech, 1999.
[6] Norbert A. Streitz, Jorg Geissler, and Torsten Holmenr. Roomware for cooperative buildings: Inte-
grated Design of Architectural Spaces and Information Spaces. In In N. A. Streitz, S. Konomi, H. J.
Burkhardt, editor, LNCS 1370, Proceedings of the First International Workshop on Cooperative Build-
ings, Darmstadt, February 1999.
[7] M. Weiser. Some Computer Science Issue in Ubiquitous Computing. Communications on the ACM,
(36(7): 75-84), July 1993.
[8] M. Weiser and J. S. Brown. Design Calm Technology. PowerGrid Journal, July 1996.

View publication stats