Document Sample
Brussels,-17-January-2002 Powered By Docstoc
					                                            INTERNAL REFLECTION GROUP



The EU software market for 2000 was estimated at 170 B Euro. This figure includes
packaged software, embedded software, software traded as network services and non
traded software (that is, software developed by end-user organisations).

The EU packaged software market was valued 50 B Euro by EITO in its 2001 report.

There are more than 5 B embedded systems sold every year in Europe. According to
ARTES (Sweden) and DARPA (US), 98% of all microprocessors are in embedded
applications and most of them have a control or monitoring function. The value of this
market is difficult to define by a precise figure but all analyst consider it, at least, as big as
the packaged software market.

To trade software as a networked service is a new domain closely related to the
outsourcing trend. OVUM predicts for 2004 a global market of 46 B Euro of which 9 will
be on the EU.

To estimate the value of the non traded software, OECD made a study on customers
allocation of software spending. The conclusion was that for Europe 59% of the software
is developed internally, 24% is spent on packaged software and 16% on outside

In addition to these 170 B Euro once could also include in the software market the value
of the support services such as software integration, consultancy, training and
maintenance, as indicated in figure 1.


                                                                                                 Support services
                         Million Euro

                                                           15771                         34762
                                        60000     14837                      31247               Operation management
                                                                    27917                        Implementation
                                        40000     21878                                          Consulting
                                        20000              25260

                                                  6604     7724     9012     10421       11904
                                                 1998     1999     2000     2001     2002

 Figure 1: Expenditures in software-related services in the EU (source EITO 2001).

It is also worth mentioning that in general the growth on individual software productivity
lags significantly behind the growth of needs and the improvement in performance/cost
ratio of hardware as indicated in figure 2.

                                                  Source : ITEA
           Figure 2: Expected growth in software productivity and needs

Finally to say, the Business Software Alliance estimated for the EU software market a
piracy rate of 40%.

Europe is weak in generic software, the package offering being confined to niche areas
such as ERP (Enterprise Resource Planning) but Europe is acknowledged as being strong
in end-user developed software, in the emerging services market and in embedded


Research domain: Embedded Systems (EMS)

One of the characteristics of the Information Society is the extensive integration of
embedded components in systems used in various application areas like
telecommunication, automotive, manufacturing, medical applications and e-commerce. In
most cases embedded components are real-time systems that continuously interact with
other systems and the physical world. The role of the embedded software is to
“configure” the computing device so as to meet physical requirements. In non-embedded
software, where physical properties are secondary, functional composition is the focus of
the software technology.

The main areas of today’s research are:

Activity 1 : Control theory and control engineering

New challenges include networked controls and fault tolerance in adaptive, resource
aware and self- sustained modes, as well as structured (mostly hierarchical and hybrid)
control of complex systems. Complexity is in terms of order and number of nodes, or
nonlinear dynamics, or in terms of uncertainty and unpredictability versus autonomy and
scale. Implementation of such systems becomes an IT intensive process. Europe is strong
in fundamental research in control engineering and modeling.

Activity 2: Real time systems

New challenges include highly composable hard real time systems, universal theories and
design methodologies for integration of hard and soft real time and handling of event and
time constraints as well all non functional requirements. For hard real-time systems there
is a need to validate behaviours prior to deployment. There is therefore a growing need
for new models and analysing tools. This in particular includes formal representations and
resource requirement models and schedulability analysis. Assurance of real-time systems
are cruical as real-time systems normally all are safety critical. This calls for the use of
rigorous techniques and integrated validation tools for ultra-stable embedded systems.

Activity 3 : Fault identification, management and fault tolerant control

Newer multi-disciplinary domain which complements the regulatory and optimizing
control systems. Fault management in embedded systems requires the propagation of
effects of physical and logical faults across physical and information system boundaries.
The current software technology, which builds systems in layers of abstraction,
completely loses traceability of the effects of physical faults on system behavior.
Therefore new design technologies are required together with intelligent diagnostics
software based on, for example, agent technology.

Activity 4: Domain specific platforms

Research is needed to approach embedded systems from a Systems Engineering point of
view, that is, how to realize for a specific sector (biotechnology, transport, navigation,
telecommunications, environment or home) an integrated platform which has to solve the
complete problem. The platform have to include integrated real time and control
components for design, for implementation (run-time) and for maintenance and advanced
cognitive and perception systems (vision, speech and sensing) with on-line requirements,
both distributed and centralized. Safety, optimization and monitoring are key issues.

Research domain: Distributed Systems

The research topics on distributed systems have evolved from parallel processing and
computing clusters where to share scarce hardware was the main driver to wide area
networked distributed systems where hardware (computing, storage and bandwidth) is
considered as a commodity and the driver is to add intelligence by dynamic resource
allocation and management and/or by cognitive techniques for object and event

The main areas of today’s research are:

Activity 1: Agent Technologies

Agents are autonomous pieces of software who act on behalf of humans. For instance,
and agent who will find and book the cheapest plain ticket for you to go tomorrow to
Barcelona. Agent architectures, programming and communication languages exist but
there is not a standardised way to build agent systems. Any way it is considered this will
be solved by market forces. The main research on this topic is on adding intelligence to
software agents, what the community calls “social behaviour”, that is, agents have to have
negotiation and compromise skills, agents have to distinguish between trusted agents and
non-trusted, etc.

Activity 2: Middleware

Middleware is a distributed software layer which abstracts the complexity and
heterogeneity of the underlying network technologies, computer architectures, operating
systems and programming languages to ease programming and managing distributed
applications. Today’s middlewares are object oriented. The most popular middleware is
OMG’s CORBA but Sun and Microsoft are also active in this domain and market forces
will decide which approach is to prevail. Researchers in this domain look for more
intelligent middlewares. Examples are: A.- Event-based middleware (to build applications
that must monitor and react to changes in their environment), B.- Reflective middleware
(for systems that inspect and adapt at run time).

Activity 3: Distributed Databases

Networked databases are mainly object databases which main issues are data integrity
management, transparency, replication and concurrency. Many systems exist in the
commercial market and interoperability and metadata management are main issues but
researchers in this domain are moving to more intelligent structures as: 1.- Mobile
Databases which going beyond the client server domain query execution is partly done on
the mobile devices or 2.- Organic Databases (which can be embedded in hardware
appliances and provide adaptative, self-healing and stable data store-recall functionality to
its environment).

Activity 4: Fully distributed systems with no (or very limited) central management and

The most known systems are: A.- File Sharing Services as Napster and Gnutella B.- P2P
(Peer-to-Peer) systems where a considerable computing power is created by aggregating a
big number of small computer resources provided by their owners to be used by the
community when idle. P2P goes beyond the client-server model (all peers have the
capacity to be both client and server). Related to these are the GRID technologies created
by the research community to solve complex problems one single machine (even the most
powerful) can not handle. But P2P differs from GRID: in P2P there is not an overall
structure were every machine plugs-in, in grids the logical peer for each machine is not a
peer but the grid. Additionally the GRID is supposed to link supercomputing centres but
the P2P elements are supposed to be PCs, PDAs or mobile phones. This is a new domain
researchers look into new ways of building up self organising systems composed of
millions of small autonomous devices.

Activity 5: Application services provision

This is the provision of software applications as managed services over VPN (Virtual
Private Networks) or over WAN (Wide Area Networks). Market research organisations
(as Gartner, OVUM, etc.) predicted incredible growth figures for this market for 2001 but
the market has not followed because, in between other factors, much research is need to
provide secure, scalable and affordable ASP systems.

Other activities

- Cluster Computing: this is moving from the traditional inter-link of local computers to
wireless-clusters and to clusters of small remote devices.

- Computer supported co-operative work. This is a domain driven by the user’s needs
mainly in the aeronautics, car and building industry on what is called distributed
- Programmable networks. To support service interworking and users management in the
a context of many network and service providers new techniques to reconfigure users
assigned resources dynamically and to handle service hand-over are needed. These
include active and programmable networks were the network’s nodes execute user
packets in order to reconfigure to their needs.

Research domain: Software Technologies

The 22nd International Conference on Software Engineering resulted in a book: The
Future of Software Engineering which presents a summary of the state-of-the-art. The
fields covered are: Software Process, Requirements Engineering, Reverse Engineering,
Testing, Software Maintenance and Evolution, Software Architecture, Object-oriented
Modeling, Software Engineering and Middleware, Software Analysis, Formal
Specification, Mathematical Foundations of Software Engineering, Software Reliability
and Dependability, Software Engineering and Performance, Software Engineering for
Real-Time, Software Engineering for Safety, Software Engineering for Security, Software
Engineering for Mobility, Software Engineering Tools and Environments, Software
Configuration Management, Databases in Software Engineering, Software Engineering
and the Internet, Software Economics, Empirical Studies of Software Engineering,
Software Metrics, Software Engineering Education. . In each of these fields a roadmap
setting out the key challenges, is presented alongside an outline of the ways in which such
challenges might be met This information
forms a valuable source of information for setting the research agenda.

The Ambient Intelligence vision includes software components for nearly all devices. To
make this a reality software production processes (development, maintenance, support,
etc.) are to be improved to increase individual productivity without compromising on
reliability. Software technology research faces today three main challenges:

(1)    produce more effective development and operation environments

(2)    produce a new generation of enabling technologies

(3)    extend and enrich the conceptual foundations.

The main areas of today's research are:

Activity 1: Architectures, Component-based Software Engineering (CBSE)

Architectural research is particularly focussing on treating functional and non-functional
requirements at the design stage. To enable reasoning at the architectural level on issues
like composability, quality, scalability, reliability or dependability will be crucial for
effective development environments. Research in this field includes formal specifications
and modelling. Attention is moving to real-time aspects, self-correction, self adaptation
and dynamic evolvability and reconfigurability (of distributed) software systems.

The approach to building systems out of components is being more and more widely
accepted as presenting major advantages in terms of lowering systems development
costs, speeding up development, and increasing quality. The growing importance of
distributed Web applications and the many possibilities of ambient and ubiquitous
computing now make the component approach to be perceived as the only way to make
such applications possible. Research is still needed in this field (evolution of systems
made of components and dynamic setup of such systems), but issues of introduction

(business models, liability, standard representations) will also need attention at other

Important work is being done in the architectural context to develop means to predict
dependability of complex systems.

Activity 2: Distributed development, EUD and Process Improvement

The needs for collaborative software development by geographically distributed teams
and individual developers is increasing. This creates specific constraints on software
development processes, related to collaborative work, composability and quality. Classic
process technologies will not be adequate in a distributed environment.

More generally, the classic software development processes, whose introduction has
obtained significant support in earlier Community programmes, is currently more and
more replaced by agile programming methodologies. Other development enabling
techniques will however in particular focus on effective use of architectures and
modeling for the actual development of the programming code.

Software configuration, maintenance and management is an important aspect of the
software process. The main aim is not only to keep a software system running but to
develop new design techniques in order to allow a software system to evolve keeping
pace with the changes not only on the underlying technology but also on the user needs.
Self-monitoring techniques are to be embedded into it.

Another topic that could lead to increased software productivity and gains increasing
attention is End User Development (EUD). It aims at delegating design or programming
capabilities to the users through easy-to-use "software development mechanisms"
embedded within the applications. EUD challenges the problem of dynamic adaptation of
software and systems functions to end-user's needs that are evolving or depending on the
context of use. EUD consists of a set of techniques (such as programming by
demonstration, macro recording, scripting, spreadsheets programming, etc.) empowering
end-users to instruct their applications, without the need to understand complex software
engineering or programming concepts.

Activity 3: Information Management and related Knowledge Engineering issues

Software tools and components for support of knowledge bases (ontologies) modeling,
design, development and reuse with syntactical verification and semantic construction.
Semi-manual and automatic techniques for multimedia, semi-structured or textual content
tagging and labelling, and for structuring, indexing, enriching, annotating large
information data sets with metadata. Personalised access to search, retrieval and delivery
of knowledge and information extracted from large and distributed data repositories
(including spatio-temporal and geographic data) are additional features.

Activity 4: Service creation environments, End user services

Work focuses on interoperability, standards, platforms and tools. The objective is to
provide generic infra-structures, tools and building blocks that enable wide-scale and
efficient creation and deployment of services. The different requirements for classes of
services in vertical areas (e.g. knowledge, content or commercial services) will require
specific building blocks and tools. Tools and technologies for low-cost design and

engineering of human-computer interaction in services will be needed to support the
acceptance of such services.


Research domain: Embedded Systems

The European scientific community in embedded systems has organized an international
workshop in preparation for FWP6, Vienna March 26, 2001 and is strongly
participating in the EMSOFT conference (1st one in Lake Tahoe, US , Oct 2001, 2nd one
planned in Grenoble, Oct 2002) .

At scientific level DG INFSO/E1 is in contact with the IFAC strategy group and EUCA
(European Union Control Association). Relevant activities are also planned at IFAC 2002
world congress, the two Mediterranean Control Conferences 2002 and 2003 and the
European Control Conference 2003.

In 2002, the 3-annual IFAC world Congress will take place, after 12 years, back in
Europe (Barcelona) and the presidency will continue to be in European hands

There are more than 220 labs in Europe. To build synergies, DG INFSO/E1 is funding the
open thematic network ARTIST which has today 22 teams.

The Real time systems is a 25 years old domain. The main international body IEEE

Europe is leading the world in the scientific work and related tools for hard real time
design (in particular synchronous systems) and in static analysis of real time systems. In
the area of real time communication systems, and in electricity distribution real time
controls, Europe is leading the efforts of some international standardization bodies.

In Fault identification, management and fault tolerant control, Europe has teams
internationally recognized as the University of Hull (Prof. Ron Patton) and Technical
University of Vienna (Prof. H. Kopetz).

In the EUREKA framework member states are co-operating in the industry-driven ITEA
programme. Its aim is to promote the development of embedded and distributed software.
Founding companies include Alcatel, Barco, Bosch, Bull, DaimlerChrysler, Italtel, Nokia,
Philips Electronics, Siemens, Thales, and Thomson multimedia. There is a regular formal
contact between IST and ITEA. An elaborate and rich Technology Roadmap on Software
Intensive      Systems      was       produced    early     2001       http://www.itea-

Research domain: Distributed Systems

There is a well established research community. IEEE Computer Society has a very active
group since the beginning of the 80’s. Recently they decided to publish an on-line
magazine which gives a good overview of the
domain. ICDCS 2002 (International Conference on Distributed Computing Systems) is in
its 22nd issue and will be held this July in Vienna IFIP
(International Federation for Information Process) and ACM (Association for Computing

Machinery) are both also active in distributed systems (e.g. last November they jointly
organised the International Conference on Distributed Systems Platforms in Heidelberg).

On Agent Technologies, Agent-Link is a thematic network running in DG INFSO/E1 Its 131 members (of which 30% are from industry) include
nearly all organisations doing research in this domain in Europe. AgentLink prepares
courses, publishes a news letter, keeps a Web server, supports conferences and have set
up 5 Special Interest Groups (SIGs) to produce a roadmap by the end of the year which
will be written by Christine Guilfoyle (who wrote the well known OVUM report on agent

On Middleware EUTIST- AMI is a cluster of middleware projects running in DG
INFSO/E1. 19 organisations are members (of which 25% are from industry). Membership
will be extended after 8th Call. LogOn Technology Transfer which is the organisation
representing the OMG (Object Management Group) in Europe is also member. OMG
(, is an open membership, not-for-profit consortium that produces and
maintains computer industry specifications of middleware for interoperable enterprise

On Distributed Databases, there are many EU funded projects but there isn’t neither a
thematic network nor a cluster. The Database Systems Group of University of Trier
(Germany) is maintaining a good list of world-wide research groups in this domain

On Fully distributed systems with no (or very limited) central management and control,
DG INFSO/E1 have a few projects coming form the last calls but the constituency is not
yet defined as to set up a thematic network. A good starting point is the Global
Computing initiative launched by DG INFSO/F1.
An important market initiative is the JXTA project launched by Sun

The Application services provision constituency can be found around the ASP Industry
Consortium at and at (ASP
Konsortium). This domain was called for in the 7th Call as a CPA. Selected projects are
now under negotiation. It will be a cluster, running in DG INFSO/C4 of about 20
organisations in between them the ASP Industry Consortium.

Research domain: Software Technologies

There is a well established global software engineering research (SWE) community
meeting yearly in the International Conference on Software Engineering. Information on
upcoming conferences of ICSE, see: .
European fora are ICSE: and ESEC-FSE, European SE
conference: ; . Other more
specific communities include software architecture ( ),
software product lines ( ),
modeling and formal methods (EMMSAD'01), dependable systems (ISSRE).

On Component Based Software Engineering, DGINFSO/E2 is running CBSEnet, a
thematic network which will contribute to the elaboration of an R&D roadmap and will
contribute to speed up the transfer of new technologies to the industry.

Other networks within this field, are: 26436 WG-ECUA, European COTS user working
group, 28754 ESERNET, Experimental Software Engineering
Network and 28767 QUESTION-HOW, Quality Engineering
Solutions via Tools, Information and Outreach for the new Highly enriched offerings
from W3C

On Information Management and related Knowledge Engineering issues the constituency
is quite diverse but its of particular interest the area of ontology engineering .

The End User Development is relatively new field building multi-disciplinary teams
(HCI, Visual Languages, Software Engineering, CBSE, Artificial Intelligence, Machine
Learning, Cognitive Interfaces, Requirement Engineering, Computer-Supported Co-
operative Work, Software Agents, Participatory Design, Evolutionary Software
Development, etc.). Around 20 universities in Europe are currently developing research,
in between them, Université de Paris-Sud, Fraunhofer-Gesellschaft                  FIT
(Forschungszentrum Informationstechnik), University of Bari and CNR/CNUCE.

With respect to service creation environments the focus should be on research and
industrial organisations that can play a role in the global work on interoperability and
standard protocols and formats. The most important platforms are W3C (,
the World Wide Web Consortium responsible for Web standardisation activities (e.g.
HTML, XML, …), OMG (, the Object Management Group producing and
maintaining computer industry specifications for interoperable enterprise applications, the
Open Group (, for standards process and infrastructure
services, OASIS (, the Organization of Associated Salespeople in the
Southwest, Inc. (OASIS representing the manufacturing, sales and distribution side of the
giftware industry and WWRF (, the Wireless
World Research Forum.



DG INFSO/E2 is responsible for the Working Group established by the ISTAG to
provide advise on Software and Systems Technology for the realisation of ISTAG vision
on Ambient Intelligence. Recommendations and priorities for the IST Work Programme
2003 are expected in April 2002, whilst a final report is foreseen for June 2002. It is
important to notice that the outcomes of this working group will be important not only for
the 3 research domains described in this concrete bullet point of the Specific Programme
but also for all other parts of the Programme concerned with the development of Software
Systems for the Ambient Intelligence vision.

Additionally to ISTAG members the work within WG9 involves a number of experts in
the field, therefore its creation is to be understood as part of the constituency building
exercise. The second meeting of these experts will take place on 7 March and this
opportunity will be used to provide information on the progress in the development of

The final report of WG9 will also be a good opportunity for organising an open workshop
to extend the constituency and stimulate further thinking on the priorities proposed. This
would preferably be in Sept / Oct. 2002.

Free and Open Source Community

Significant discussion is ongoing in the developers world on the deployment of free /
open source software. Although certain research might specifically relate to this
development methodology (e.g. distributed development, without central control), the
discussion is particularly relevant from a point of view of industrial strategy. Specific
areas of software development, like standards, interoperability protocols, development
platforms or other middleware, might benefit considerable from this movement.

DG INFSO/E2 is closely following this community. A significant set of projects is
funded and these are informally clustered. A consultation meeting on perspectives in the
FP6 time-frame was held in May 2001 with a good representation of international open
source software players. Report at:
These links with the free and open source movement will continue as part of the actions
to address the constituency not only for this concrete bullet point of the Specific
Programme but also for all other parts of the Programme concerned with this movement.

Actions within Thematic Networks and Clusters

For the 3 research domains, nearly all activities quoted before have (or have now under
negotiation) an open thematic network, which should be the main mechanism for
addressing the specific constituencies. These networks publish News Letters, organise
workshops and create consensus among its members.

Additionally, the text of call 8 under AL IV.3.2 allows for submission of accompanying
actions to prepare work under FP6, in the whole field of Technologies and Engineering
for Software, Systems and Services.

Fostering co-operation with Industry Associations and Open Fora

Some Industry associations, consortia and open fora are directly involved in our activities.
In all other cases there are well established links. This makes the participation of
Commission officials in domain specific conferences quite easy. Additionally, in the
past, it has been very effective to publish Commission information (e.g. the launch of a
Call) on their News Letters. This co-operation will continue as part of the actions to
address the constituency.

Fostering co-operation with US and National programmes

Co-operation an joint events with US constituencies and agencies e.g. NSF and DARPA
is ongoing, particularly on the field of embedded systems and on dependability.

Co-operation with national programmes on the field of software technologies is ongoing
with France, where the government supports a thematic networks co-managed by the
Ministries of Industry and Research on software technology and its generic applications. and

On embedded systems co-operation is established with The Netherlands where a national
institute has been set up and with Austria where a national research programme on the
this domain is funded.

News Letter on Software and Services of Unit E2


Events at which the constituency can be addressed and for which presentations on FP6 and
Software technologies are arranged with E2.
General conferences
 SNPD '02 - 3rd ACIS International Conference on
  Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed
  Computing - June 26-28, 2002, Madrid,

    Invitation received..

 SEKE '02: The Fourteenth International Conference on SOFTWARE

    Invitation received

  ENGINEERING TRACK, Dortmund, Germany, September 4th - 6th, 2002

    Invitation received

E2 network and working group workshops
 Workshops of Working Group 9 of ISTAG. With selected group of experts on 7
  March and in June.
  An open seminar on the final report can be organised in sep/ Oct 2002. To be

 2nd ESERNET scientific workshop, October. Place and exact date to be determined.
  Presentration agreed

 ECUA workshop will be held on 6-7 May 2002 in DELTA, Hoersholm (closed to
  Copenhagen), Denmark. Presentation agreed.

    (Poznan Technical University will organise the ECUA Dissemination event in
    Poland on 20 June 2002, IT Ltd. will have it on the second half of June in Hungary,
    while Sofia University has not decided the date yet (it will most likely be in June or
    September as July and August are seen as quiet months for conferences in Europe)).

 Workshop CBSE Net, expected Oct/Nov 2002. Presentation ok.


Shared By:
Description: Brussels,-17-January-2002