Federated management of information for TeleCARE

Hamideh Afsarmanesh, Victor Guevara-Masis, L.O. Hertzberger

University of Amsterdam, Faculty of Science

Kruislaan 403, 1098 SJ Amsterdam, The Netherlands

Abstract. Distributed information management plays a fundamental role within

the base infrastructure supporting the elderly care domain. Specificities of this

domain include the autonomy and independence of its involved actors, the

critical data that is handled about individuals, and the variety of

hardware/software resources supporting the elderly care environment. A

federated information management system coping with these requirements is

designed and integrated as a core component of a mobile agent-based

infrastructure, to support collaborative networks for elderly care. Functionalities

for: federated schema management, federated query processing, HW/SW

resource management, specification and enforcement of visibility/access rights

to data and resources, and an ontology-based automatic schema generation

facility are introduced, and their implementation details are briefly discussed.

1 Introduction

A technologically advanced elderly care environment is a highly distributed

collaborative network, composed of heterogeneous and autonomous nodes. Each node

in the network (e.g. a care providing center, a leisure providing organization, a

supermarket/shop in the neighborhood, a relative or a friend of the elderly person,

etc.) is in one way or another interested in supporting the elderly person with his/her

independent living style (see Fig. 1). Therefore, each node individually plays an

assisting role, and further in collaboration with the other nodes in the network,

provides certain organized support and services to the elderly.

The collaboration among different active organizations (e.g. care centers and

leisure centers nodes) in the network forms a so called Virtual Organization (VO) [4],

while the collaboration among active individuals (e.g. relatives and friends nodes) in

the network forms a Virtual Community (VC) [10].

To develop an advanced elderly care environment, in addition to the Internet

representing the communication facility for the network, some special base

requirements must be satisfied for reliability, safety, and privacy of the information

exchanged through the network. Furthermore, a number of other advanced

capabilities and features are required, where among others, the remote

operation/access to HW/SW resources (e.g. to home-devices and support-services),

interoperation of heterogeneous information sources and software systems, tele-

monitoring, federated management of the information and cataloging the HW/SW

Afsarmanesh H., Guevara-Masis V. and Hertzberger L. (2004).

Federated management of information for TeleCARE.

In Proceedings of the 1st International Workshop on Tele-Care and Collaborative Virtual Communities in Elderly Care, pages 49-62

DOI: 10.5220/0002683800490062

 SciTePress

resources information, error recovery, and cooperative problem-solving for tele-

assistance services, can be emphasized.

Leisure site

Care center 3

Emergency

mobile site

Mobility for elderly

Relative’s

monitoring site 2

Virtual shop 1

Special

doctor site

Elderly home sites

Care center 2

Virtual shop 2

Relative’s

monitoring site 1

Police StationPolice Station

Care center 1

Fig. 1 – The TeleCARE scenario for virtual elderly support environment

The IST TeleCARE project aims at design and development of a base

infrastructure to support elderly care environments, addressing many of the above

challenges. A scenario for TeleCARE is depicted in Fig. 1. Proper handling and

access to all related information within the elderly care environment, is of vital

importance, and thus the information management constitutes a key element of the

TeleCARE infrastructure. Considering the independence and autonomy of the

network nodes, a main challenge here is the organization, management, and provision

of retrieval facilities for both the heterogeneous data stored at the nodes, as well as

the information regarding its HW/SW resources (i.e. devices and services) at each

node within the network, while preserving their access rights and authorization.

Furthermore, considering the incremental development of the support services to be

made available within the elderly care network, developers of such services require

assistance to make their services interoperate with other existing systems and

resources. One problematic issue here is that in order for services to be connected to

the elderly care network, their developers must structure and store all their data within

the network’s database. To avoid the need for expertise in database modeling on the

network’s database, it should be sufficient for developers to use an editor through

which they can define their data by its ontology, that shall in turn be automatically

translated into proper database structures and get stored in the network database.

This paper first briefly addresses the TeleCARE platform architecture and its main

elements. It then presents the three main components developed in TeleCARE project

to support and facilitate the management of all the information related to elderly care

environment. These components include: FIMA – Federated Information

Management component, RCAM – Resource Catalogue Management, and DOSG –

Dynamic Ontology-based data Structure Generation. Although the above components

are currently developed to benefit the area of elderly care networks, these components

are generic enough, so that at least a very large part of the designed and developed

components can be applied to any advanced emerging collaborative network.

2 TeleCARE as a tele-assistance platform

The overall goal of the IST 5FP TeleCARE project is the design and development of a

configurable framework solution for tele-supervision and tele-assistance, to support

the elderly. The proposed solution has been seen as complementary to other

initiatives, focused on the integration of elderly in the society to reduce their isolation

[14]. The TeleCARE solution benefits from the merge of a number of technologies

and paradigms in order to provide an open architecture supporting seamless future

expansion. In specific, it is based on the integration of: (i) multi-agent systems

(MAS), including both stationary and mobile intelligent agents, (ii) federated

database systems, (iii) Secure communication, and (iv) the services that are likely to

be offered by the emerging ubiquitous computing and intelligent home appliances.

In a nutshell, the core horizontal platform developed for TeleCARE provides the

MAS, mobility, safe communication, and the federated information management

services. The TeleCARE consortium further develops some vertical services on top of

this platform, including status monitoring, as well as other forms of assistance such as

agenda reminders, entertainment services, time bank, and a few base services

supporting virtual communities, web-accesses, and specialized elderly user-interface.

2.1 The TeleCARE reference architecture

The reference architecture for TeleCARE nodes providing cooperation/federation

among different nodes of the elderly care network is depicted in Fig. 2. The main

components of this architecture are briefly addressed in this section.

Vertical Services Level

Core MAS Platform Level

External Enabler Level

Virtual Community

Support

Web-access

Support

Specialized

Interfaces

Time Bank

service

Status

Monitoring

service

Enter-

tainment

service

Agenda

service

Platform Manager

Basic Multi-

Agent Platform

Resource Catalogue Mgnt.

Inter-platform mobility

Federated Information

Management

Safe Communication

Infrastructure

Device Abstraction Layer

Inter-agent Communication

Persistence

Support

Ontology

Management

System

Inference

Engine

Agent

Exit

Control

Agent

Reception &

Registration

Agent

Factory

Device/Vertical Service Registry

Resource

Managers

Federated

Query

Processor

Ontology-based

Schema

Generator

Specialized

Components

TeleCARE Basic Platform

Fig. 2 - The TeleCARE platform reference architecture

The designed architecture of TeleCARE node is composed of a three-level

platform. The External Enabler Level provides support for the external

communication of the TeleCARE node, and the interfaces with external resources.

The Core MAS Platform Level is the major component of this architecture, and

includes essential support for software agents. Finally, the Vertical Services Level is

the open component where a variety of services can be gradually added to the node.

External Enabler Level. This level supports the remote communication with other

nodes and provides interfacing mechanisms to the external devices. This level

compromises two segments:

• Safe communications infrastructure - providing safe communications and

supporting both secure and reliable agent/messages passing among different nodes.

• Device abstraction layer - interfacing the sensors, monitoring devices, and other

hardware (home appliances, environment controllers, etc.) to the TeleCARE

environment.

Core MAS Platform Level. The platform level is the main component of the

environment and offers fundamental services for agents as well as for their

interactions. These services include the creation, launching, reception, user

authentication, access rights verification, and execution of stationary and mobile

agents. The main modules at this level include:

• Basic Multi-Agent Platform - provides the essential multi-agent support, and it is

based on Aglets framework [8] with the following extensions:

i. Ontology management system - The Protégé 2000 [13] is used in the

platform for the definition of the ontologies.

ii. Inference engine - For intelligent agent interpretation using a Prolog

interpreter.

iii. Persistence support - For basic recovery mechanisms.

• Inter-platform mobility - extension to the basic MAS platform to support

generalized mobility of agents, including agent security mechanisms. This module

includes the Agent Reception & Registration component, and the Agent Exit

Control component, for administration of stationary and mobile agents.

• Inter-agent communication - extension to support credentials and coordination of

agent communication, independent of the agent location.

• Federated information management - supporting the management of information

at TeleCARE nodes and providing the infrastructure for i) flexible processing of

federated queries, ii) data structure generation based on ontological definitions and,

iii) preserving information privacy through access rights management. This

component is developed using Java with free/open source software, it is built on

top of SAP DB relational database system [11], and Castor data binding

middleware for Java [6].

• Resource catalogue management - managing the catalogue of resources, and

registering the descriptions of all device and vertical services available at the site

as well as their access rights.

•

Agent factory - supporting the creation / specification of new agents.

• Platform manager - configuration and specification of the operating conditions of

the platform at each site, including user administration and node management.

Vertical Services Level. The applications and vertical services level focuses on the

actual support for the elderly (which require specialized user interfaces), care

providers, and elderly relatives (assuming that they are able to interact with normal

computer interfaces). Further, it compromises the two layers.

• Base services - set of base services on top of the horizontal infrastructure that

each provides specific support to other value-added services.

i. Virtual Community Support - To support the management of Virtual

Community for the elderly care environment

ii. Specialized interfaces for elderly - Suitable computer interfaces for the

use of elderly person

iii. Web-access Support - Web-based mechanisms to interface with the

TeleCARE environment.

• Vertical Services - A number of specialized vertical services are implemented as

specific TeleCARE applications, including a VC-based time bank, an elderly

status monitoring service, an elderly agenda reminder service, and an elderly

entertainment service.

2.2 Management of information in TeleCARE

The analysis of information management requirements for the TeleCARE network

has identified both the modeling and functionality required to be supported local to

each node, as well as for the information exchange/integration and necessary

interoperation among the sites. Based on the analysis of these requirements, the

necessity of three main components were identified that together support both the

management of all information related to TeleCARE network. These components,

namely FIMA, RCAM and DOSG are briefly described in this paper.

3 FIMA - Federated Information Management component

The Federated Information MAnagement (FIMA) component of TeleCARE supports

applications that may require variety of data models and large numbers of users and

agents accessing and retrieving its data, while supporting the pre-defined visibility

rights to physically distributed and heterogeneous data. The federated database

architecture of FIMA does not require any centralization of data or control and thus

supports flexibility and extensibility aspects required for future use of TeleCARE

system. The database architecture can also support a variety of application

architectures that may be used for development of different vertical services for

TeleCARE, including both the client/server and the agent-based systems. The

database repository of FIMA is developed using the SAP DB as the base. The SAP

DB provides an open source and freeware DBMS, and was selected for TeleCARE

among many considered options for this purpose.

The two key functionalities offered by FIMA include:

- Federated data and Schema Management, that handles all the data and schemas

defined in the network, while supporting the definition of adequate levels of

information privacy for access by authorized agents / users.

- Federated Query Processing, which supports the collection of all necessary data

from different distributed heterogeneous and autonomous nodes through a single

query issued by the user as if all data distributed among different nodes are in fact

available at the local site.

Fig. 3 shows a high level architecture of the FIMA and its main software

components. These components, depending on their role and functionality, are all

implemented as stationary and / or mobile agents. Detailed description of all these

agents is outside the scope of this paper. Below we focus on the Federated Query

Processing of FIMA and provide details on the stationary and mobile agents

supporting this functionality, and how the query processing performance is improved

in comparison to other agent-based approaches [9].

Other

internal

components

TeleCARE node

Device

abstraction layer

-DIMA -

Data Interface Mapping Access

-AIMS -

Agent Information

Management System

FIMS Agent

-MIRA -

Mobile Information

Retrieval Agent

Base Service

Service A

Service B

Service C

Vertical

Services Level

Core MAS

Platform Level

External Enabler Level

DBMS

-IAMA -

Information

Access Manager

FIMA

M I R A

F I M S

Fig. 3 – Architecture of Federated Information MAnagement – FIMA

The processing of federated queries is a complex task, and it is briefly detailed as

follows. First, the requester sends a query (which is in high-level format) to the FIMA

interface, which generates an agent designed to handle this request. The query is then

translated considering the internal structures of the stored data, and a set of sub-

queries is established. These sub-queries are one by one assigned to mobile agents

with the proper itinerary. After this step, these mobile agents are dispatched to the

remote nodes to accomplish their mission, to perform the local query, and to send the

results back to the original node. Finally the received results are merged at the node

and returned back to the requester, see Fig. 4.

What should be noticed here is that the main goal of the federated query processor

component in FIMA is to enable TeleCARE agents and end-users to query the

authorized information, while hiding all the details about database connections, agents

creation and their traveling among nodes, and processing the data.

Below are the main agents involved in federated query processing of FIMA:

• FIMS Agent: Federated Information Management Server Agent, acting as the

FIMA interface agent

• FQP Agent: Federated Query Processor Agent, acting as the query supervisor

• MIRA Agent: Mobile Information Retrieval Agent, acting as the mobile

component, transferring the jobs to other nodes, for this process.

FQP instruction

(XML specification)

Query

execution

FIMS Agent

-Agent interface

FQP Agent

- Query Processing

MIRAs

- Access to remote

information

Requester

agent

Fig. 4 - Main components of the federated query processing

3.1 Federated Information Management Server Agent

The Federated Information Management Server Agent (FIMS Agent) manages the

interface to access the information in FIMA. It must be continuously available and

running. It supports multi-users, and thus can fulfill requests from numerous agents

simultaneously, that may have different purposes other than just executing a single

query. FIMS Agent does not handle however all the query processing related

operations, otherwise it will too overloaded. Whenever FIMS receives a request for a

federated query, it generates another agent (FQP) in a different execution thread, thus

allowing it to maintain its primary operation. This mechanism provides the highest

performance for query processing, since the new FQP agent will be focused only on

the task of performing the query.

When the FIMS Agent creates an FQP Agent to supervise the processing of a

federated query, it also provides the identifier of this FQP Agent to the requester. As a

result, from that point on, all query operations are bound to that FQP Agent, and the

FIMS Agent is freed from the responsibility of following the federated query

execution.

3.2 Federated Query Processor Agent

The Federated Query Processor Agent (FQP Agent) is at the heart of the federated

data processing in FIMA. It implements several advanced techniques, particularly

useful in the distributed TeleCARE environment. Some of the used mechanisms for

federated query processing of FIMA are also integral to the TeleCARE infrastructure

(e.g., the multi-agent and the Java object oriented programming environments).

A number of techniques are used to improve the performance of the query

processing, for example: (1) special multi-thread processing, (2) simultaneous

execution of several queries, and (3) reduction of communication costs by reducing

the size (i.e. content) of the mobile agents involved in FQP. All these mechanisms

focus on the internal operations for the processing of federated queries. The number

of these internal operations is large and they are grouped in several task categories,

which can be summarized as follows:

• Query translation: the query that arrives in high level functional format in

XML, is first translated into internal handling structures.

• MIRA creation: depending on the type of federated query, appropriate Mobile

Information Retrieval Agents (MIRAs) are created, e.g. if the query type is

"parallel", then multiple MIRAs will be created, one per target node.

• Query decomposition: the original query is divided into a number of sub-

queries according to the number of target nodes, and these sub-queries are

assigned to the corresponding MIRA agents.

• MIRA transmission: Each MIRA is sent to a remote node, carrying the

corresponding sub-query.

• Query evaluation: The MIRA agent performs the communication MIRA-to-

FIMS Agent of the remote node, in order to execute and retrieve the requested

information from that node.

• Result transmission: The MIRA transmits to the FQP Agent the information

resulted from the sub-query.

• Information merge: Once all results arrive from the MIRA to the FQP Agent,

the FQP merges the sub-results, and sends the final results to the requester.

• Resource release: When the execution of the query completes, the requester can

agree to release resources generated by the FQP Agent, disposing all the MIRA

agents involved in the query evaluation as well as the FQP itself. Note that

disposing the FQP Agent at any stage of the query execution will effectively

close the processing of the federated query.

3.3 Mobile Information Retrieval Agent

The Mobile Information Retrieval Agent (MIRA) is a mobile agent that transmits the

federated query to other nodes. Being a software agent, it also guarantees the

possibility of combining intelligent decision making with the information retrieval

tasks. Therefore, it can support a range of federated queries, for example:

• the case of gathering information from several specific nodes at once and

merging the results at the originating node,

• the case of a query that “crawls” from one node to another to search all nodes,

e.g. retrieving all possible answers, or to find the best answer, or,

• the case of a query searching all nodes one by one, trying to satisfy certain

condition. One example is the case of finding the first possible answer, or to find

a satisfactory answer (e.g. by sending the answer to the originator and waiting

for a “satisfaction” response from the originator node, in order to decide either to

continue the search or to quit), or even the case of finding a specific number of

answers and then quit.

The FQP Agent creates MIRA agents and their handling is completely transparent

from the requester. Clearly, from the requester point of view, the proper execution of

the query and its results is what really matters, and not how the query mechanism was

implemented. This transparency noticeably reduces the system complexity, since the

TeleCARE application designers and developers are not concerned about internal

details of processing mechanism that they need to invoke.

As part of the strategy to enforce the visibility levels and access rights on the

information, FIMS Agent will also "borrow (from the TeleCARE platform system)

and check the credentials" of the requester agent for creating the FQP Agent. The

FQP agent in turn uses those credentials to create authorized MIRA agents. In

general, this strategy is used in FIMA to validate the access rights to the information

for requesters, no matter if the requester is local or remote. Section 3.5 refers more on

visibility levels and access rights.

3.4 Processing of query types

The federated query processing mechanism of FIMA supports access/retrieval of data

from multiple TeleCARE nodes, as such the data can be retrieved either from the

same or different remote nodes. Three types of federated queries are supported in

FIMA, that allow retrieval from remote data stored in different nodes on the network.

• Parallel query type, where the performance in speed is the key consideration.

• Serial query type, where the optimization of resource usage is the focus.

• Sequential query type, that requires interactivity with the requester to control

the information-processing overhead, see Fig. 5 below.

One advantage of providing different types of query/access methods is that for

instance the requester (for instance the vertical service that accesses some data from

the network database) can choose among the three options, to control the general

performance and overhead of the process and thus easily optimizing the performance

of the federated queries for specific purposes.

Mobile Agent

partial

results

PC 1

PC 3

PC 2

query

Requester

Agent

FQP Agent

Fig. 5 - Sequential query type in FIMA

Furthermore, the design of the architecture of FIMA carefully considers necessary

data / agent traffic among the TeleCARE nodes. Therefore, two communication

mechanisms between the nodes are properly supported, namely (1) the inter-agent

message passing and (2) the agent mobility, are respectively considered for

supporting the cases of “information push” and “information pull” among the agents:

• Information push: A simple TeleCARE information exchange case is

considered where a care center site requires the periodic sending of the sensed

data from the home sites. In this case, the data collected at the home site is

pushed from the home site to the care center. The “push” action is performed

as “messages” sent from one agent to another.

• Information pull: To illustrate the information pull, assume that an elderly

wishes to plan a special fun activity in his/her community, and starts this

planning through an “entertainment service” at home. The corresponding

“elderly entertainment service” at the home site inquires some relevant data to

be collected from the Virtual Community. Then, a MIRA agent, that may

contain a parallel, serial, or sequential federated query (depending on the kind

of request, if it is to all, or for instance specific number of people), can pull

some names from one or more leisure centers.

3.5 Visibility levels in TeleCARE

In federated information management networks, different autonomous nodes can have

different visibility levels and access rights on other nodes’ information. Thus, every

node in the federation can decide what part of its local information should be

available to each member in the federation [1, 3, 12].

In other federated database environments the approach for visibility levels is either

based on individual export schema definitions on the local schema for every external

“user”, or based on the definition of a complete hierarchy of export schemas such as

in [7]. However, due the highly dynamic nature of the TeleCARE environment,

where users and nodes are added and removed regularly, a different approach is

adopted for defining the visibility levels. This approach is based on the credential of

every agent, and specifically on the agent type, that also represents the role of the user

generating the agent.

4 DOSG - Dynamic Ontology-based data Structure Generation

Typically, for building large systems and applications, the assistance of a database

expert is required to define the structure for concepts and entities of the environment,

namely the database schema. The DOSG component supports and assists both the

TeleCARE component developers as well as its service developer, with their direct

definition and modification of database schemas, for the data that needs to be

processed by their code, while eliminating the need for database expertise. Namely,

DOSG provides facilities for dynamic and automatic definition/modification of

database schemas, so that they can be automatically stored into the database. As such,

for this purpose the service developers of TeleCARE, can simply use the user friendly

interface provided for the ontology system “Protégé” to provide their data structure

definitions.

The main focus of DOSG in TeleCARE is transforming the ontology definition

provided for some information, into the underlying information management model,

(based on the relational database system) as well as the Java objects specification.

DOSG provides a highly innovative mechanism to leverage object knowledge model

in ways which vertical service developers can use to store, retrieve and manipulate

information seamlessly through the federated information management layer of the

TeleCARE platform.

DOSG is designed as a plug-in to Protégé. It extends Protégé’s ontology editor

with an interface that allows users to parameterize the automatic data structure

generation. DOSG benefits from the integrated Protégé environment by gathering

online input related of conceptual schema, while allowing customization of some

parameters for this generation process through the DOSG interface. The

implementation of DOSG is in Java and, it also uses free / open source software, in

specific, Castor is used to produce the two mapping definitions [6], while Xerces is

applied for the development of XML Schema [2]. As shown in Fig. 6, based on the

ontological definitions provided by users, the DOSG tool automatically generates five

different outputs, namely:

• RDBMS Schema with the appropriate SQL script for relational databases,

• Java classes providing the source code of the data structures,

• XML Schema with the specification for proper handling of XML documents,

• Object-relational mapping containing the mappings that governs the

conversion between Java classes and the database system, and

• XML mapping that defines the translation between the Java classes and XML.

Ontology Manager

Protégé 2000

DOSG

Data Structure

Generation

XML Schema

XSD file

RDBMS Schema

SQL Script

Java classes

Source code

Vertical

Service

Application

Data Object

mapping

XML mapping

JAR

package

<XML>

<. . .>

</XML>

Fig. 6 – Output format for Dynamic Ontology-based data Structure Generation - DOSG

5 RCAM - Resources Catalog Management component

A complementary module to FIMA is the Resource CAtalog Management (RCAM)

component. RCAM provides definition of the resource model, supports automated

resource management, and enables TeleCARE service developers to define, search,

and modify specific details of resources available through the TeleCARE

environment. Resource descriptions in RCAM are based on widely accepted

standards, in order to allow current and future devices (e.g. house hold appliances)

and/or emerging vertical value-added services to be easier added to the TeleCARE

platform. All hardware devices and software services in TeleCARE are treated as

resources. Basically, RCAM acts like a registry for all resources, their internal service

descriptions, and interfaces. Namely, for every resource of the TeleCARE

environment, RCAM manages three types of information:

a. the catalogue entry representing a definition of the resource,

b. the entries for resource’s internal services definitions and,

c. the access rights to the resource

In order to support the current and future devices and emerging vertical services,

TeleCARE resource definitions in RCAM are based on widely accepted standards.

The hardware device definitions are based on the UPnP (Universal Plug and Play)

[15] specification, while the software vertical service definitions are based on Web

Service Definition Language (WSDL) specification [16]. Furthermore, RCAM

resources definition has been extended to better support users’ access rights to

resources, based on agent identification, part of the TeleCARE passport definition [5].

The RCAM Agent provides basic operations on the TeleCARE resource model.

These operations can be grouped into the following categories: Resource

advertisement and publishing, Resource discovery, Resource access rights

management, see Fig. 7

Resource

Broker

Resource Catalogue

Resource

Requester

Request Description

Advertisement /

Publishing

Resource service

Information

Discovery

Invocation /

Execution

Description

Resource

Provider

- Device and Vertical Services-

Resource Description

Fig. 7 - Resource handling operations

The suggested TeleCARE resource model involves the following actors: resource

provider, resource broker, and resource requester. RCAM Agent acts as an

automated resource broker. This resource broker provides a searchable (catalogue)

repository of resource definitions, through which resource providers can

advertise/publish the functionality of their resources. Additionally, resource

requesters search for appropriate resource services and obtain the necessary

information to use them.

Furthermore, RCAM can store information about the access rights to TeleCARE

resources, based on the TeleCARE passport definition. As such, for every resource,

related information regarding the Agent-type, User-role, and User-id of its authorized

users can be stored. Thus, every time that a Resource Manager agent receives a

request to access its resource, the Resource Manager can first access RCAM to

validate the authorization of this access for the specific requester. Therefore, usage of

RCAM, properly secures the access and usage of the vertical services and devices.

6 Conclusions

A federated information management approach offers suitable mechanisms to cope

with the required flexibility, heterogeneity, autonomy, and privacy requirements for

information handled within collaborative networks for elderly care. The combination

of this approach with a mobile agent platform has proved to be an effective approach

to develop a flexible infrastructure supporting a large variety of TeleCARE services.

The developed prototype system supports information interoperability between

agent-based systems, contributing to an open plug-and-play philosophy involving a

resource variety of hardware devices and appliances, as well as software vertical

applications and services. The federated query processing in TeleCARE transparently

provides access to remote data from several nodes and supports different types of

queries. The dynamic ontology-based data structure generation facility offers

system/service developers a new level of flexibility as they can focus on modeling

their tasks at the ontology level using a user-friendly interface. Finally a modular

approach is introduced for resources (devices and services) to be integrated in

TeleCARE via the Resource Catalog Management component, and thus making it

possible for resources to be discovered and applied in the future service developments

for this environment.

References

1. Afsarmanesh H. and Camarinha-Matos L.M., "Federated Information Management for

Cooperative Virtual Organizations," presented at 8th International Conference on

Databases and Expert Systems Applications, DEXA'97, Toulouse, France, 1997.

2. Apache Software Foundation, "Xerces2 Java Parser," http://xml.apache.org/xerces2-

j/index.html, 2004.

3. Busse S., Kutsche R.-D., Leser U., and Weber H., "Federated Information Systems:

Concepts, Terminology and Architectures," Computergestutzte

InformationssystemeCIS, Technische Universitat Berlin, Berlin Forschungsberichte des

Fachbereichs Informatik 99-9, 1999.

4. Camarinha-Matos L.M. and Afsarmanesh H., "The virtual enterprise concept, in

Infrastructures for Virtual Enterprises," Camarinha-Matos, L.M. and Afsarmanesh, H.,

Eds.: Kluwer Academic Publishers, pp. 3-14, 1999.

5. Camarinha-Matos L.M. and Afsarmanesh H., "A multi-agent based infrastructure to

support virtual communities in elderly care," To appear in International Journal of

Networking and Virtual Organizations. 2004.

6. ExoLab Group, "Castor," http://www.castor.org/, 2004.

7. Garita C., Afsarmanesh H., and Hertzberger L.O., "The PRODNET Federated

Information Management Approach for Virtual Enterprise Support," Journal of

Intelligent Manufacturing, 2000.

8. IBM Tokyo Research Laboratory, "Aglets Workbench," IBM Japan,

http://www.trl.ibm.com/aglets/, 2002.

9. Papastavrou S., Samaras G., and Pitoura E., "Mobile agents for WWW distributed

database access," presented at ICDE, 1999.

10. Rheingold H., The virtual community: Homesteading on the electronic frontier:

Addison-Wesley, 1993.

11. SAP AG, "SAP DB," http://www.sapdb.org/, 2003.

12. Sheth A. and Larson J., "Federated Database Systems for Managing Distributed,

Heterogeneous, and Autonomous Databases," ACM Computing Surveys, vol. 22, pp.

183-236, 1990.

13. Stanford Medical Informatics, "The Protege Project," Stanford University School of

Medicine, http://protege.stanford.edu/, 2003.

14. TeleCARE, "A Multi-Agent Tele-Supervision System for Elderly Care,"

http://www.uninova.pt/~telecare, 2003.

15. UPnP Forum, "Universal Plug and Play," http://www.upnp.org/, 2002.

16. W3C Consortium, "Web Services Description Language (WSDL),"

http://www.w3.org/TR/wsdl, 2001.