Concepts and System Structures to Support Collaborating Everyday Items 

Thomas Schoch Institute for Pervasive Computing, ETH Zurich 

Abstract 

The goal of this dissertation is to contribute towards the realization of the 
vision of a world with "smart" everyday items. Smart everyday items differ 
from regular everyday items insofar as they know their whereabouts, perceive 
their environment, and are able to communicate with other smart things. For 
example, a bottle that contains temperature sensitive chemicals can monitor 
the temperature and if a certain limit is exceeded, the bottle can send an 
alarm message; and if its own temperature sensor is broken, it simply asks a 
bottle in its vicinity. Another task that a bottle might take over is less 
complex and more cost efficient: it signs in and off at the warehouse 
management systems of supply chain participants, so that they know exactly 
what inventory is currently stored. 

In the introductory first chapter of this work, this vision will be extended 
and it will be shown how it can be embedded in the more general vision of 
ubiquitous computing, which was first formulated by its pioneer Marc Weiser. 
Starting from these visions, economically practicable applications will be 
introduced in the following second chapter. One generic application, which 
will be referenced throughout this work, will be discussed more deeply. It 
consists of a supply chain application with several participants that exchange 
products. At first, these economic applications serve, especially the supply 
chain application, to enumerate the requirements that need to be fulfilled by 
systems that aim at realizing a world of smart everyday items. By this, we 
understand a software system that comprises, on the one hand, a software 
framework that provides application developers with the required abstractions 
in terms of a generic class library and, on the other hand, middleware 
services that extract and execute application-independent tasks. It is the 
assumption of this dissertation that the concepts and systems presented here 
describe and support a world of collaborating everyday items in a 
substantially better way than would be possible with current means. 

In the third chapter, the technologies are presented on which the smart thing 
systems considered in this work are based. It is about software systems, on 
the one hand, that support dynamic client/server applications, such as Jini or 
Web Services, and hardware systems, on the other hand, that serve to identify 
objects automatically, such as bar code or RFID systems. 

After these preparations, the fourth chapter introduces a model that 
describes a world of collaborating everyday items. On the highest level, the 
model differentiates between four concepts: a thing, the tag attached to it 
for identification, a representation of that thing in IT systems, and a 
service infrastructure} that allows for the coupling of the tag and its 
representation. These four concepts and their relative dependencies will be 
expanded and presented. Besides the definition of the vocabulary that is used 
throughout this work, the model serves to formulate the problems and 
challenges of an implementation of such systems precisely, so that the model 
can be used as a template for future implementation. During the work with the 
model, the following problem areas and questions arose that are also addressed 
by the model: "How can we abstract from the implementation details of the 
underlying identification technology such as bar codes or RFID?", "How must 
an identifier be structured?", "Which kind of location model is 
appropriate?", "How does the service infrastructure realize the coupling 
between a thing and its representation?", "Which other services can be 
provided by the service infrastructure?", "What is the minimum functionality 
that has to be provided by a representation?", "What are the possible 
options for migrating a representation?" and "How is the interaction between 
applications and representations realized?". 

The implementation of the model in three different smart thing systems as 
well as the corresponding insights, are the topics of the fifth chapter. Both 
topics can be regarded as a solution to the above mentioned problem areas. 
First, the question of how to structure applications that use one of the three 
systems will be addressed. A similar issue is how such systems and the 
applications based on them can communicate with already existing applications. 
Building on this basis, a description is given of the three systems that cover 
different aspects of the model, so that a subsequent system extends a previous 
one, as well as trying out different technologies. Some features that 
distinguish the systems are the usage of Jini or Web Services as an underlying 
client/server platform, various location models, and the option of a 
representation either migrating from one host to another in order to be 
executed as near as possible to the location of the thing, or of the 
representation being permanently located at a host where it is reachable from 
a client application at all times. 

The deployment of the smart thing systems, i.e. the execution of the 
application and the middleware service on computers, as well as the 
application development, are evaluated in the sixth chapter. This evaluation 
relies, on the one hand, on the supply chain application introduced in the 
second chapter as a generic example and, on the other hand, on the measurement 
of the relevant system parameters of the underlying technology. Finally, the 
results will point out whether the systems actually support application 
developers and whether they scale well enough to be deployed on a larger 
scale. 

The seventh chapter shows to what extent links or overlaps to related work 
exist, thereby suggesting the broader value of the concepts and 
implementations presented here. Mainly those projects are presented that also 
aim to realize a world of smart things, as well as work that is intended for a 
different goal but has similar subordinate problems to solve. 

In the last eighth chapter, building on the evaluation of the sixth chapter, 
the pros and cons of the model and the smart thing systems are discussed, and 
potential extensions of the model and the systems are sketched. In addition, 
adjacent research areas and their challenges are addressed.