An Evolutionary System Development Approach in A Pervasive Computing

Environment

Ruimin Liu Hongji Yang

Software Technology Research Laboratory De Software Technology Research Laboratory De
Montfort University, Leicester, UK Montfort University Leicester, UK
rliu@dmu.ac.uk hyang@dmu.ac.uk

Yangsheng Wang Wei Pan

Institute of Automation, Chinese Academy of Institute of Automation, Chinese Academy of
Sciences, Beijing, China Sciences, Beijing, China
yangsheng.wang@mail.ia.ac.cn wpan@nlpr.ia.ac.cn

Abstract [3], but making them effectively invisible to the user [2].
Pervasive computing names the third wave in computing.
Mark Weiser described a hypothetical world in which
humans and computers were seamlessly united in a
seminal article a decade ago; he called this vision
Ubiquitous Computing [5, 6] (which now is also called
Pervasive Computing). After that, technology has
advanced along many dimensions, especially in hardware
progress and wireless communication technologies. A
number of leading technological organisations are
exploring Pervasive Computing. But it is far from
Weiser’s vision become reality. Pervasive Computing will
be the future. Pervasive computing will be a fertile source
of challenging research problems in computer systems for
many years to come. But until today, there is less attention
to the system evolution in pervasive computing
environment. This paper discusses the current
development of Pervasive Computing and technical
Figure 1. Major Trend in Computing [3]
challenges in deploying Pervasive Computing systems.
First, a model for evolution of pervasive Information
Pervasive computing is fundamentally characterised by
technology infrastructure of pervasive information is
the connection of things in the world with computation.
presented which shows how the existing technologies
Pervasive computing just might help to free the minds
solutions available can be used to develop a pervasive
from unnecessary work, and connect people to the
information infrastructure. Then a generic architecture of
fundamental challenge that humans have always had: to
pervasive computing system is given. Based these, an
understand the patterns in the universe and ourselves with
evolutionary system design approach is suggested.
them. Weiser heralded a future of ubiquitous and invisible
computing in a now often referenced article in Scientific
American in which he stated: “The most profound
1. Introduction technologies are those that disappear. They weave
themselves into the fabric of everyday life until they are
The concept of pervasive computing emerged in the indistinguishable from it.” [5] When articulated, this was a
early 1990s by Mark Weiser. Pervasive computing is the vision too far ahead of its time -- the hardware technology
method of enhancing computer use by making many needed to achieve it simply did not exist.
computers available throughout the physical environment A pervasive computing system should be pervasive,
embedded, nomadic, adaptable, powerful, yet efficient,

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE
 intentional and eternal. It must everywhere, with every organisations represent a broad communal effort to make
portal reaching into the same information base; live in our pervasive computing a reality, there are many useful
world, sensing and affecting it; allow users and utilities can be used, both in enterprise and in research
computations to move around freely, according to their field, include ubiquitous devices (PDAs, smart devices,
needs; provide flexibility and spontaneity, in response to e.g.), software for pervasive computing environment (Java,
changes in user requirements and operating conditions; operating systems, middleware components, e.g.) and
free itself from constraints imposed by bounded hardware infrastructure ( wireless networks…). Based all these
resources, addressing instead system constraints imposed contributions, we attempt an evolutionary design approach
by user demands and available power or communication in pervasive computing environment. The key progress
bandwidth; enable people to name services and software includes component extraction from existing pervasive
objects by intent, as opposed to by address; never shut computing systems and seamless components integration
down or reboot; components may come and go in [11] with the Jini Technologies.
response to demand, errors, and upgrades [14]. The research project on which this paper is based
After a decade of technology progress (include new proposes the method to extract useful components from
technologies—such as the Global Positioning System existing pervasive computing systems (such as Oxygen,
(GPS[16], smart cards, and radio frequency identification Aura, Portalano, e.g.) according to the new requirements.
(RFID) tags—and social developments such as the Using the Jini technologies, all the components, including
increasingly widespread acceptance of video surveillance the existing components (hardware components and
in public places), many critical elements of pervasive software components) and the components extracted from
computing are now viable commercial products, such as: existing pervasive computing system, can be seamless
handheld and wearable computers; wireless LANs; and integrated into an evolutionary system. The difficult
devices to sense and control appliances. problems lie in component extraction, component
A number of leading technological organisations are synthesis and the application of Jini technologies.
exploring Pervasive computing. Xerox's Palo Alto The structure of this paper is as follows. Section 2 is
Research Centre (PARC), for example, has been working the overview of the proposed approach. Section 3 is the
on pervasive computing applications since the 1980s. model of evolution of pervasive information infrastructure.
Although new technologies are emerging, the most crucial Section 4 gives a generic architecture of pervasive
objective is not, necessarily, to develop new technologies. computing system. Section 5 is the framework for
IBM's project Planet Blue, for example, is largely focused component identification and integration. Section 6 gives
on finding ways to integrate existing technologies with a blueprints of the future work.
wireless infrastructure. Carnegie Mellon University's
Human Computer Interaction Institute (HCII) is working 2. An Approach to Evolutionary System
on similar research in their Project Aura, whose stated Design
goal is "to provide each user with an invisible halo of
computing and information services that persists Pervasive Computing will be the future. Pervasive
regardless of location." The Massachusetts Institute of computing will be a fertile source of challenging research
Technology (MIT) has a project called Oxygen. MIT problems in computer systems for many years to come.
named their project after that substance because they But until today, there is less attention to the system
envision a future of pervasive computing devices as freely evolution in pervasive computing environment. The
available and easily accessible as oxygen is today. system evolution in pervasive computing environment
Pervasive Computing will be the future [15]. Pervasive includes many aspects, such as the evolution of the
computing will be a fertile source of challenging research information technology infrastructure, the evolution of
problems in computer systems for many years to come. software system, etc. Pervasive computing system also
There are some major research challenges in pervasive has the same characteristics as other computing systems,
computing environment [8], such as software architecture, though it has many more requirements than prior
networking challenge, component interaction, user computing system, such as distributed system and mobile
interface integration, e.g. [7] But until today, there is less computing system. So the development of pervasive
attention to the system evolution in pervasive computing computing is also an evolutionary process.
environment. Pervasive computing system also has the Because many research organisations represent a broad
same characteristics as other computing systems, though it communal effort to make pervasive computing a reality,
has many more requirements than prior computing system, there are many useful utilities can be used, both in
such as distributed system and mobile computing system. enterprise and in academic field, include ubiquitous
So the development of pervasive computing is also an devices (PDAs, smart devices, e.g.), software for
evolutionary process [1]. Because many research pervasive computing environment (Java, operating

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE
 systems, middleware components, e.g.) and infrastructure 3. Evolution of Pervasive Information
( wireless networks…). Based all these contributions, we Technology Infrastructure
attempt an evolutionary design approach in pervasive
computing environment.
The concept of pervasive computing is a paradigm shift
from a traditional computing point of view. So pervasive
computing should be modelled as a gradual evolution
from an existing information technology infrastructure
[10]. The evolution to pervasive computing is
characterised by a marked increase in mobility,
embeddedness, information technology integration, device
heterogeneity and the communication capability from an
infrastructure point of view. These basic characteristics
and other attributes—primarily derivatives in nature—
such as adaptivity, smartness (intelligent infrastructure),
localised scalability, uneven conditioning and context
sensitivity are the basis for the evolution of pervasive
infrastructure [12].

Figure 2. Proposed Evolutionary System

Design Approach

The approach consists of three major steps shown in

Fig 2: Requirements, Component Recovery, Component
Integration and Deployment.
Based the generic architecture of pervasive computing
system mentioned in Section 4 of this paper, the first step
of evolutionary pervasive computing system design is
requirement analysis. To the given scenario, the
requirement of system can be found out.
Component-based system design is the major approach Figure 3. Evolution Stages of Pervasive
in the construction of pervasive computing system. Computing Infrastructure
Pervasive computing has its own characteristics: a
pervasive computing system should be pervasive, There are four stages in pervasive information
embedded, nomadic, adaptable, powerful, yet efficient, technology infrastructure evolution process as shown in
intentional and eternal. So, a set of rule to do system Figure 2.
evolution should be established before the component • Stage 1 is traditional computing infrastructure, which
extraction. We have concluded some rules for system is computing PC centric, no organized introduction of
evolution in pervasive computing environment, such as mobility in the information technology infrastructure, low
component taxonomy. level of automation, focuses on integration of business
The key of the research project is the method to extract systems and workflow automation, use of information
useful components from existing pervasive computing technology for ERP, SCM, CRM and organisation value
systems (such as Oxygen, Aura, Portalano, e.g.) according consolidation.
to the new requirements. Using the Jini technologies, all • Stage 2 is mobile computing infrastructure which is
the components, including the existing components adoption of mobile technology and mobile information
(hardware components and software components) and the access integrated into traditional computing infrastructure.
components extracted from existing pervasive computing Mobile applications primarily for field, force
system, can be seamless integrated into an evolutionary professionals and basic PIM, wireless capability missing
system. The difficult problems lie in component extraction, or limited to cellular access. Automation is still low.
component synthesis and the application of Jini Increased focus on mobilising the workflow.
technologies. This will be expressed in Section 5. • Stage 3 is integration of wireless infrastructure which
is adoption of wireless technology, both cellular and
WLANs. In this stage, widespread use of mobile devices
across the organisation in new user segments like

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE
 knowledge workers with wireless email access, infrastructure like displays, radio Ids (e.g., RFIDs) and
introduction of platforms for transaction management, radio devices such as those based on Bluetooth, infrared,
devices management, content transcoding and mobile etc.) and a high level of technology integration (software
security, supporting for ad hoc networking, automation integration of different user and environmental contexts,
widespread across the organisation. In this stage, integrating user and world model with other applications,
emphasis on real-time management, real-time handing of integrating voice and other IO mechanisms). This
CRM, SCM, inventory control and dispatch. transition is primarily software intensive and will involve
• Stage 4 is pervasive computing infrastructure. intelligent software layering with a view to maximising
Introduction of pervasive infrastructure capabilities like the benefits offered by the underlying pervasive
“context sensing”, “smart space”, efficiency management computing infrastructure. This transition will also see
systems for devices heterogeneity, “location awareness”, additional pervasive computing capabilities based on the
new short range wireless technology such as Bluetooth, a experiences gained in the previous stage. High levels of
strong communication infrastructure, always-on application integration will be involved. The modification
connectivity, content management and personalisation of existing application base to include pervasive
platforms, bandwidth management systems, voice characteristics may also be involved.
platforms, substantial increase in the level of automation,
a software fabric encompassing the entire pervasive 4. A General Architecture of Pervasive
computing infrastructure providing infrastructure Computing System
abstraction and providing all pervasive computing
capabilities as services to all other applications, system The generic architecture of pervasive computing based
operates in a very proactive manner and lots of on the evolutionary view of pervasive computing
transactions happen invisibly based on user and infrastructure is showed in Figure 3. The purpose of this
environment contexts and implicit input sensing, new architecture is to providing a structured evolution
applications can be rolled out very fast without any roadmap to systematically build a pervasive computing
concern about the underlying infrastructure details, infrastructure while accounting for the existing technology
increasing automation of transactions, focus on [9].
collaboration, automation, real-time management and This is a plug-and-play architecture for a pervasive
real-time networking, infrastructure support for real-time computing infrastructure which can be increased by
decision making. adding new technology components and by means of
From stage 1 to 2, transition typically characterised by integration. The overall architecture is based on an
new technology adoption (mobile devices, existing mobile computing infrastructure with a truly
synchronisation systems) and a relatively low level of integrated communication layer. The communication layer
technology integration. Primary running costs are related consists of technologies like WLAN, Bluetooth, infrared
to procuring and maintaining devices. Fixed infrastructure and cellular technologies like
costs involve synchronisation software, device GSM/GPRS/UMTS/CDMA1X and provides wireless
management software and security infrastructure. connectivity as a service provided by the infrastructure.
From stage 2 to 3, transition typically characterised by The communication layer should be extensible so as to
new technology adoption (wireless LAN setup, wireless allow both indoor and outdoor (cellular) wireless access
modems (WLAN, cellular) for mobile devices (laptops, depending on the context. A wireless capability is
PDAs) ) and a medium level of technology integration. absolutely fundamental in the evolution from a mobile
Platforms for transaction management, device computing infrastructure to a pervasive computing
management, content transcoding and mobile security, etc. infrastructure. The backend infrastructure for supply chain
need to be introduced. Adding real-time transactions to management, ERP, business intelligence, vertical
the previously synchronisation based mobile applications integration, etc., remains as it exists in the
communication model. Wireless access to e-mail and mobile computing infrastructure context [12].
intranet access can lead to improved productivity. The core components consist of services such as
Mobility across the enterprise can lead to more service/ resource discovery, security, transactions,
collaborative work environments. The infrastructure more notification, query engines, storage services, replication
conducive towards handling ad hoc structures Wireless services, messaging engine and session management, etc.
capability gained at this stage is a primary requirement in The plug-n-play components typically reside at the
the next stages of the pervasive computing evolution. periphery of the pervasive information infrastructure
From stage 3 to 4, transition typically characterised by along with other components like mobile devices. As an
a relatively low-level new technology adoption (context example, plug-n-play infrastructure components could be
sensing including location sensing, smart space those related to context sensing like location information

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE
 (location sensors-transmitters, tags and other location
sensing devices), infrared beaming, radio ID tags such as
RFIDs, tracking capabilities and other sensors such as
temperature, lighting, etc.

Figure 5. A Framework for Component

Identification and Integration

• JAIW is a component integration tool, can discover,

match, and assemble components to complete an
application. It’s also responsible for recognising what
resources are available and noticing when there is a
resource change in the user environment. It’s based the
Jini technology.

Figure 4. A Generic Architecture of the Pervasive

Computing System

Whenever a new plug-n-play component gets added,

say, indoor location sensing, only the corresponding
virtual of the concerned objects need to be updated with a
new context definition. The system can now provide
enhanced services by virtue of the fact that the smartness
of the system (smartness is a fundamental characteristic of
pervasiveness), and hence pervasiveness, has increased. Figure 6. Overview of Jini Technology in JAIW
This concept not only allows for rapid service creation but
also helps to evolve the pervasive computing Figure 5 shows the overview of the Jini technology
infrastructure without impacting the applications that the approach. RFID systems are connected to event drivers
user interacts with [12]. (EDs), which generate enter and leave events from
periodical tag scans. The EDs act as producers for the
5. A Framework for Component counterpart event service (VCES). The VCES delivers
Identification and Integration events to the counterpart manager (VCM), and specific
counterparts. The VCM acts as an execution environment
The evolutionary pervasive computing system design is for counterparts. Upon the first sighting of a tagged object
component-based [4]. We proposed a framework based on or location, it consults the counterpart repository (VCR) to
Jini technology in this process. This is a part of our obtain counterpart executables for the tag or location.
approach in pervasive computing system evolution. Counterparts register with the look-up service (LUS), so
Figure 4 gives the proposed framework in evolutionary that cooperating counterparts can find each other. The
pervasive computing system design. There are different artefact memory (AM) acts as a place for persistently
components in this framework [6]. storing and retrieving counterpart state and event histories.
• Task Manager maintains a representation of tasks, Small amounts of state can also be stored in the tag
and a mapping between context and user preferences. memory by sending appropriate store events to the VCES
• Context Observer interprets the physical context of [13].
the user and is responsible for identifying such things as
the user’s location, focus of attention.

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE
 6. Conclusion and Future Work Conference on Software Architecture, Kluwer Academic
Publishers, pp. 29-43, August, 2002.
[5] M. Weiser, "The Computer for the Twenty-First Century,"
Pervasive computing will be a fertile source of Scientific American, pp. 94-10, September, 1991.
challenging research problems in computer systems for [6] H. Yang, X. Liu and H. Zedan: “Abstraction: A Key Notion
many years to come. Solving these problems will require for Reverse Engineering in a System Reengineering Approach”,
us to broaden our discourse on some topics, and to revisit Journal of Software Maintenance 12(4): 197-228, 2000.
long-standing design assumptions in others. We will also [7] M. Satyanarayanan, “Pervasive Computing: Vision and
have to address research challenges in areas outside Challenges”, IEEE Personal Communications, 2001.
computer systems. These areas include human-computer [8] G. Borriello, “Key Challenges in Communication for
interaction (especially multi-modal interactions and Ubiquitous Computing”, 50th Anniversary Issue, IEEE
human-centric hardware designs), software agents (with Communication, 2002.
[9] R. Grimm, T. Anderson, B. Bershad and D. Wetherall, “A
specific relevance to high-level proactive behaviour), and System Architecture for Pervasive Computing”, Proceedings of
expert systems and artificial intelligence (particularly in the 9th ACM SIGOPS European Workshop, pages 177-182,
the areas of decision making and planning). Capabilities Denmark, September, 2000.
from these areas will need to be integrated with the kinds [10] M. Burger, “The Path Towards Pervasive Computing -- A
of computer systems capabilities discussed in this paper. Network Approach”, O’REILLY Emerging Technology
Pervasive computing will thus be the crucible in which Conference, February, 2002.
many disjoint areas of research are fused. [11]A. Kameas, S. Bellis and I. Mavrommati, “An Architecture
We should be fully aware that attaining it would that Treats Everyday Objects as Communicating Tangible
require tremendous creativity and effort by many people, Components”, Proceedings of the First IEEE International
Conference on Pervasive Computing and Communications--
sustained over many years. The early decades of the 21st PerCom’03, March, 2003.
century will be a period of excitement and ferment, as [12] P. Gupta and D. Moitra, “Evolving a Pervasive IT
new hardware technologies converge with research Infrastructure: A Technology Integration Approach”, Personal
progress on the many fundamental problems. Like the and Ubiquitous Computing, Volume 8, Issue 1, ACM, February,
Frontier of the American West in the early 19th century, 2004.
pervasive computing offers new beginnings for the [13] K. Romer, T. Schoch, F. Mattern and T. Dubendorfer,
adventurous and the restless — a rich open space where “Smart Identification Frameworks for Ubiquitous Computing
the rules have yet to be written and the borders yet to be Applications”, Proceedings of the First IEEE International
drawn. Conference on Pervasive Computing and Communications--
PerCom’03, March, 2003.
Our future work can be addressed as follow. First, [14] R. Liu and H. Yang, “Software Evolution in Ubiquitous
further investigation into the infrastructure of pervasive Computing”, Proceedings of CACSUK 2003, Manchester,
computing is required, especially in middleware England, September, 2003.
infrastructure (like some operating systems for pervasive [15] U. Hansmann, L. Merk, M. S. Nicklous and T. Stober,
computing environment) and wireless Web. Second, how Pervasive Computing Handbook, Springer, 2001.
to combine the existing technologies of software [16] M. Rahnema, “Overview of the GSM System and Protocol
reengineering to the evolution of pervasive computing Architecture,” IEEE Communications, April, 1993.
systems and make the evolution available are our
assumption of future work. We propose to build a model
for pervasive computing system evolution based on the
approach we mentioned in this paper.

Reference
[1] H. Yang and M. Ward, Successful Evolution of Software
Systems, Artech House, Boston, USA & London, UK, ISBN:
1580533493, January, 2003.
[2] M. Weiser, "Some Computer Science Problems in
Ubiquitous Computing," Communications of the ACM, July
1993.
[3] M. Weiser, “Ubiquitous Computing”, IEEE Computer,
October, 1993.
[4] J. P. Sousa and D. Garlan, “Aura: An Architectural
Framework for User Mobilityin Ubiquitous Computing
Environments”, Proceedings of the 3rd Working IEEE/IFIP

Proceedings of the 2004 International Conference on Cyberworlds (CW’04)

0-7695-2140-1/04 $ 20.00 IEEE