Jussi Kiljander, Matti Eteläperä, Janne Takalo-Mattila, Juha-Pekka Soininen
VTT Technical Research Centre of Finland, Kaitoväylä 1, Oulu, Finland
Kari Keinänen
VTT Technical Research Centre of Finland, Vuoriniementie 3, Espoo, Finland
Keywords: Smart Environment, Smart-M3, File Transfer, Semantic Interoperability, Ontology.
Abstract: Smart Environment is a physical place where different kinds of devices interact meaningfully with each
other to assist and support us in our everyday life. A key requirement for enabling these kinds of smart
systems in physical spaces is the ability to share files autonomously between various devices in the
environment. In this paper a novel solution for autonomous file sharing between heterogeneous devices is
presented. Here autonomous means that the devices interact with each other seamlessly and the
heterogeneity of devices and services providing file sharing is hidden from the end user. Our approach is
based on presenting information about the files and file sharing protocols by using ontologies and we utilize
Smart-M3 as the platform for sharing semantic information between devices in a physical space. To
demonstrate our approach for autonomous file sharing in practice we have implemented a meeting
application to various mobile platforms.
Our environment is full of heterogeneous electronic
devices such as computers, mobile phones and home
appliances, for example. Pervasive and Ubiquitous
Computing are computing paradigms aiming to
enable Smart Environments where these devices
interact meaningfully with each other in order to
provide useful services for people in the
environment (Weiser, 1991). Files are a significant
part of persistent storage in modern computer
systems and file sharing is an integral part of many
inter-device applications. These applications include
the Web and email, for example. It is also apparent
that many Smart Environment applications would
require devices to share files such as videos, music,
images, documents and even software with each
other. File sharing is therefore also an important part
of interaction between ubiquitous devices and a
common solution for file sharing in Smart
Environment is needed.
A typical scenario of file transfer in Smart
Environment can be divided into two parts:
discovery of file sharing services and execution of a
selected functionality of the service. The discovery
functionality is usually provided by some Service
Oriented Architecture (SOA) solution such as
Universal Plug and Play (UPnP), Web Services or
Common Object Request Broker Architecture
(CORBA), for example (UPnP forum, 2008), (W3C,
2002), (OMG, 2002). Also other non-SOA solution
for service discovery exist e.g. Bluetooth Service
Discovery Profile (SDP) and the Zeroconf standard
(Cheshire and Steinberg, 2005).
In addition to discovering the file services, a
device must be able to interact with the services to
execute specific file transfer functionality such as to
store, modify or request a file, for example. For this
purpose many standard file sharing protocols based
on the traditional client/server paradigm exists.
These protocols include the File Transfer Protocol
(FTP), Secure Copy (SCP) and Apple Filing
Protocol (AFP), just to name a few. Additionally
numerous SOA solutions have their own methods
for describing services that provide file sharing
capabilities. As an example UPnP requires a priori
standardization of the service interfaces were as the
Web Services and CORBA, for example, utilize
technologies such as Web Service Description
Language (WSDL) and Interface Definition
Kiljander J., Eteläperä M., Takalo-Mattila J., Soininen J. and Keinänen K..
DOI: 10.5220/0003362501910196
In Proceedings of the 1st International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS-2011), pages
ISBN: 978-989-8425-48-5
2011 SCITEPRESS (Science and Technology Publications, Lda.)
language (IDL) for describing the service interfaces
in a machine interpretable format.
As can be seen there are many service solutions
and protocols available that provide file sharing
capabilities for devices. In addition the variety of
communication technologies such as the ZigBee,
Bluetooth and Wireless Local Area Network
(WLAN) creates their own challenges for
autonomous file sharing in Smart Environments.
With a large amount of different communication
technologies, file sharing technologies and SOA-
based file sharing services it can be a difficult both
to find all available file transfer services and to
interact with the services to perform the required
operation. Basically this requires that each device
sharing a file must use all discovery mechanisms it
supports for finding the available file services and
then publish the file to each service it is capable to
interact with. Additionally to interoperate with
devices that do not utilize SOA solutions the devices
are also required to advertise the file with each
communication technology it supports. It is apparent
that this creates a lot of overhead both to the code
size of applications and to the communication
between the devices.
In this paper we propose a novel solution for
autonomous file sharing with heterogeneous devices,
file transfer methods and SOA solutions by utilizing
an interoperability platform called Smart-M3 on top
of these existing solutions. As presented by (Liuha,
Soininen, Otaloea, 2010) the Smart-M3 achieves
semantic interoperability between devices of Smart
Environments by providing architecture for utilizing
ontology-based information presentation in a
physical space context. In our approach we present a
common ontology for defining information about the
available files, file transfer protocols and file
services in semantic form and publish this
information available to all devices in the Smart
Environment. This enables devices sharing the
common ontology to negotiate about the possible
solutions for file sharing and then select the
appropriate method for transferring the file.
Smart-M3 is an independent information level
interoperability architecture that can be implemented
on top of any communication or service level
solution. The fundamental idea of Smart-M3 is to
utilize the Semantic Web ideas of ontologies and
semantic interoperability in a physical space context
(Berners-Lee, Hendler, Lassila, 2001). In addition to
the ontology-based interoperability model Smart-M3
defines a functional architecture that specifies the
methods for accessing the semantic information in a
physical space. The Smart-M3 functional
architecture consists of Knowledge Processors (KP)
and Semantic Information Brokers (SIB). Fig. 1
presents the Smart-M3 concept including the core
elements and their relations.
Figure 1: Smart-M3 Functional Architecture.
The Smart-M3 functional architecture is based
on publish/subscribe paradigm and therefore it
enables reactive event-based communication
between KP and SIB entities. SIB is the information
repository of the Smart Space and it provides
operations for modifying and requesting the
semantic information. KPs form the actual
application for the end user by utilizing the SIB
service for sharing semantic information with each
other. In order to hide the complexity of both
information presentation and communication from
the application logic, the KP is divided into KP
Interface (KPI) and use case logic parts. KPs that
interact with each other via SIB are interoperable if
they have a common ontology that specifies the
concepts and relationships between the concepts in a
given domain.
Smart Space Access Protocol (SSAP) is the
communication protocol of Smart-M3 and it defines
the rules and syntax of operations for accessing the
semantic information in the SIB. The SSAP
specifies operations such as insert, remove and
update for modifying the semantic information as
well as various formats of query and subscribe
operations. To ensure that the content of the SIB
does not change during the execution of any
operation each SSAP transaction is executed in an
atomic fashion. In addition to these operations the
SSAP contains join and leave operations that are
used to provide access control and authentication.
The information access and security in Smart-M3-
based systems is further described in (Suomalainen,
Hyttinen and Tarvainen, 2010).
PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems
In Smart-M3 the Resource Description
Framework (RDF) and RDF-Schema (RDFS)
technologies are the backbone of the ontology-based
interoperability model (W3C, 2004). RDF is a
family of W3C specifications for modelling
metadata in form of subject, predicate and object
triples. Because RDF triples are similar to basic
sentences of a human language, it is a natural way to
make statements about information with certain
properties and relationships. RDFS is an extensible
knowledge presentation language that defines
vocabularies for building simple ontologies using
RDF. The current reference implementation of the
SIB provides three languages for querying (and
subscribing to) the RDF data: Triple, SPARQL and
Wilbur Query Language (WQL) (W3C, 2008),
(Lassila, 2002).
In our approach we divide the challenge for
autonomous file sharing into three distinct levels:
communication, service and information level. Fig. 2
presents a logical view of our approach.
Figure 2: Logical view of approach for autonomous file
sharing in Smart Environments.
The communication level covers the Open
System Interconnection (OSI) model layers from L1
to L4. It provides the basic functionality required for
transmitting bytes of data between devices.
The role of the service level is to provide
technologies for devices to share services with each
other. The Service level covers existing SOA
solutions and protocols. In the file sharing context
these protocols include the FTP, SCP, Bluetooth’s
File Transfer Profile and BitTorrent, for example. In
addition there are SOA solution such as UPnP, Web
Services, OSGi and NoTA that have their own
services for providing file transfer capabilities
(OSGi Alliance, 2005), (NoTA World, 2007).
Information level is the highest level in our
approach. The role of the information level is to
make information of devices available and the
semantics of the information interpretable for other
devices in the environment. We propose Smart-M3
to be used in the information level for sharing
semantic information required in the file sharing
In our approach necessary information about
files, transfer protocols and services is modelled as a
common ontology and the SIB is used as a common
search extent of this semantic information. This way
the devices are able to discover the available files in
a given environment and then select the most
appropriate file transfer technology for transferring
the file between devices.
We utilize RDFS vocabulary as the ontology
language in our approach. This is because the
current implementation of the SIB with WQL
supports RDFS level reasoning (also the
owl:sameAs and owl:InverseFunctionalProperty
features of the Web Ontology Language (OWL) are
supported) so the RDFS is the most practical choice.
Fig. 3 illustrates the ontology of our approach.
Figure 3: File Sharing Ontology.
The File class is the base class for all files. It
contains basic properties describing the name,
format, size (in bytes) and modification date of the
file. The ImageFile and TextFile are example
subclasses of the File class that can be used to
describe the type of the file more specifically. It is of
course possible to introduce new subclasses for the
File class and for these subclasses as well. The
FileService class is the class for all file transfer
services. The ServiceInterface class specifies the
interface of services. This is done through the
ServiceFunction and FunctionParamater classes. In
addition to accessing files through SOA services, we
also developed an elegant way to request files by
interacting only on the semantic level. For this
purpose the FileRequest and SharingProtocol classes
have been defined. FileRequest class presents a
request to a file in the SIB and it has two properties.
The hasTargetFile property defines the file (instance
of the File class) to be requested and the
hasSharingProtocol property specifies the protocols
supported by the requester. SharingProtocol class is
a base class for all various protocols enabling file
sharing. The Fig. 3 also illustrates several example
subclasses for the SharingProtocol class and it is
possible to introduce new subclasses when
To achieve autonomous file sharing on the
semantic level a KP has to be both be able to
subscribe to all the available files and all file
requests to the files the KP has published into the
SIB. This way the KP will be informed when new
file has been published or a file request to his file
has been made. Because WQL hides the RDFS
reasoning from the application developer it is
practical to use WQL when performing these
operations. An abstract syntax for Wilbur value
subscription can be presented in the following way:
R = S
(n, path) (1)
where S
is the WQL subscription operation,
R denotes the result nodes, n is the start node of the
subscription and path specifies the labels
(predicates) to be traversed from the start node in the
RDF-graph. The WQL subscription that returns all
available files can be now presented as:
R = S
(File, inv(rdf:type)) (2)
where File is the URI of the File class, inv() is a
WQL specific operations that requires the path
defined by the subexpression e to be traversed in a
reverse direction and rdf:type is a RDF property
denoting that resource is an instance of a Class.
With WQL it is also possible to perform a single
subscription that returns all instances of the
SharingProtocol class and its subclasses that are
connected to the given file via hasSharingProtocol
and hasTargetFile properties. The WQL subscription
that returns all SharingProtocol instances when a
new file request is inserted to the SIB can be
presented as:
R = S
, seq(inv(‘hasTargetFile’),
where n
is the instance of the File class and
both seq() and inv() are WQL specific operations.
The operation seq(e
) traverses a sequence of n
steps of subexpressions e
. Fig. 4 illustrates
how the queries presented in (2) and (3) traverse
through a simplified RDF-graph and return all File
and SharingProtocol instance.
Figure 4: WQL value subscriptions to the instances of
SharingProtocol and File classes.
To concretize our approach a scenario that
describes how utilizing semantic interoperability
enables autonomous file sharing between devices in
a Smart Environment is presented next. In the
beginning of the scenario there is a Web Service
with FTP server capabilities available and
information about this service is presented in the
SIB. Now a device that is capable of communicating
with the service can discover it and then store a file
and publish information about the file into the SIB.
Another device that has subscribed to all the files in
the SIB will now get an indication that a new file is
available. Let’s then say that the device would need
this file but is unable to interact with Web Services
because it supports only bluetooth for transferring
files. Without the semantic interoperability the
device would not even know that there is this kind of
file available. However, now a KP in the device can
publish information (along the common ontology in
Fig. 3) that it wants to request this file. The devices
PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems
that have this file will get notification from the SIB
that the file request has been made. If either of the
two devices support bluetooth based file transfer
they can now send the file to the bluetooth address
specified by the requester.
In order to demonstrate our approach for
autonomous file sharing in practice we have
implemented a meeting application for several
heterogeneous mobile platforms. These mobile
platforms include N900, N97, iPhone and Nexus
One. TCP/IP over Wireless Local Area Network
(WLAN) is used as the communication solution and
Zeroconf protocol provides the SIB discovery in the
WLAN. Fig. 5 presents how the mobile devices used
in the demonstration interact with each other via the
Figure 5: Meeting demonstration set-up.
In addition to the file sharing ontology presented
in the section 3 two other classes have been
specified for the ontology used in the meeting
application. Meeting class presents the available
meetings in the Smart Environment and it has three
properties. The “hasName” property defines the
display name of the meeting. The “hasParticipant”
property specifies the participants of a given
instance of the Meeting class and the “hasFile”
property defines the available files in the given
meeting. Participant class represents contact
information of meeting participants. The “hasName”
and “hasPhoneNumber” properties present the name
and phone number of the participant respectively.
Fig. 6 illustrates the Meeting ontology classes and
their relations.
Figure 6: Meeting ontology for file and contact sharing.
With this simple ontology we have created a
meeting application where users can create and
participate to meetings and share their contact
information and files with each other. By
introducing new properties for ImageFile and
TextFile classes we have also demonstrated how
easy it is to create new features to applications by
mashing-up existing information. An example
scenario of the interaction between meeting KPs and
the SIB is presented next.
Let’s say that a participant of a meeting
publishes a scientific article to the meeting through
his KP and defines himself as the author the paper.
Another user subscribed to all the files of the
meeting via his KP will now be get indication that a
file has been published. The second participant is
interested in the article published by the first
participant and makes a request to the file into the
SIB. Now the KP in the first participant’s device
receives indication from the SIB and sends the file
using the file transfer capabilities the receiving KP
has. In this case plain TCP/IP was used to stream the
file data between devices.
A new feature to this application can be
developed by utilizing the properties defined for
ImageFile and TextFile classes. By performing a
WQL subscription to authors of the file the KP of
the second user can be informed when additional
information about the author is available. Let’s say a
third person now decides to take a picture of the
author and publishes this picture with the metadata
specifying that the author is in the picture into the
SIB. Assuming there is also contact information of
the first participant available the KP in the second
participant’s device can now autonomously request
the picture and inform the user that there is
additional information about the author available.
The third user can now contact the possibly
unknown author if he wants to discuss about the
article. Fig. 7 illustrates the interaction between KPs
and the SIB in the presented scenario.
Figure 7: Sequence diagram illustrating the interaction
between KPs and the SIB.
This paper presented a novel approach for
autonomous file sharing in Smart Environments. The
fundamental idea of the approach was based on
utilizing semantic interoperability platform called
Smart-M3 in file sharing context. Necessary
information about files, file transfer protocols and
file services was modeled using ontology and a
novel solution for negotiating about the possible file
sharing methods was presented. The approach was
demonstrated by implementing a meeting
application for heterogeneous mobile platforms from
various vendors. The meeting application also
illustrated the flexibility and expandability
advantages of the ontology-based information
presentations approach to create new applications
and features by mashing-up existing information.
The paper also presented how describing
information about the content of files can take the
autonomous file transfer to the next level. However,
we only scratched the surface by defining the
authors of text files and objects of images. The next
step to be done is to describe the contents of files
more specifically. This way the KPs could
automatically make more advanced searches to the
files based on the context and user preferences.
This work has been funded by the Device and
Interoperable Ecosystem (DIEM) and Open
Ubiquitous Technology (OPUTE) projects.
Berners-Lee, T., Hendler, J., Lassila, O., 2001, The
Semantic Web, Scientific American Magazine,
[online] Available at: http://www.scientificamerican.
com/article.cfm?id=the-semantic-web [Accessed 18
October 2010].
Cheshire, S., Steinberg, D., 2005. Zero Configuration
Networking, the Definitive Guide. O'Reilly Media.
Lassila, O., 2002. Taking the RDF Model Theory Out for
a Spin”, ISWC 2002, 1st International Semantic Web
Conference, 9-12 June 2002
Sardinia, Italia, pp. 307-317.
Liuha, P., Soininen, J., Otaolea, Raul, 2010. SOFIA:
Opening embedded information for smart applications.
In Embedded World 2010, Nuremberg, Germany, 2-4
March, 2010.
NoTA World Open Architecture Initiative website,
[online] Available at:
[Accessed 12 October 2010].
OMG: The OMG’s CORBA Website, [online] Available
at: <> [Accessed 12 October
OSGi Alliance website, [online] Available at: [Accessed 12
October 2010].
RDF Vocabulary Description Language 1.0: RDF Schema,
W3C Recommendation 10 February 2004, [online]
Available at:
[Accessed 14 October 2010].
SPARQL Query Language for RDF, W3C
Recommendation 15 January 2008, [online]. Available
at: [Accessed
14 October 2010].
Suomalainen, J., Hyttinen, P., Tarvainen, P., 2010. Secure
information sharing between heterogeneous embedded
devices. In Proc. ECSA 2010, pp. 205-212.
UPnP forum, 2008. UPnP Device Architecture 1.1, 136 p,
[online] Available at: <
DeviceArchitecture-v1.1.pdf> [Accessed 12 October
W3C: The W3C’s Web Services Activity website, [online]
Available at: <>
[Accessed 12 October 2010].
Weiser, M., 1991. The Computer for the 21st Century,
Scientific American, 263(3), pp. 94-104.
PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems