ALCATEL ESPACE (F) - BELGACOM (B) - CERT (F) - University of Namur (B)
- IXI (F) - OBLOG Software (P) - ORIGIN Belgium (B) - Free University
of Amsterdam (NL)
Contents
Information in this page covers the following aspects of the 2RARE
project :
A summary of the project
The
underlying technologies
The people working on the project
The main objective of the 2RARE project is to deal with the changing requirements of these applications. Therefore, the project proposes to use a new generation of Requirements Engineering (RE) techniques with the objective of providing a rigorous specification of the requirements expressed by customers and of the system's analysis resulting from these requirements. Moreover, requirements traceability techniques will be applied in order to integrate and maintain the different kinds of descriptions captured during the RE activity.
Emerging modelling techniques based on recent proposals made in ESPRIT projects (IS-CORE, ICARUS and NATURE) will be used for the handling of the two applications. They will be complemented with pre-traceability techniques (dealing with the interrelation between the requirements and their origin) and with post-traceability techniques (tracing each requirement down to system's objects).
Suppliers will provide these techniques and support their application by providing tools, training and assistance. Users will apply the techniques by themselves and assess their suitability for industrial applications concerning the estimation of impacts resulting from changes in requirements and the system's evolution through the reuse and adaptation of existing descriptions.
The total duration of the 2RARE project is one year and the effort is about 7 person-years. The consortium is made of two users (representing the major part of the effort and accompanied by appropriate associate partners) and of two suppliers from four different countries. According to its goals and expected results, 2RARE contributes to the objective of the Task 1.2 of the Software Intensive Systems Engineering subdomain in the Software Technologies domain by evaluating the benefit of integrating modelling and traceability leading-edge technologies and both suppliers and customers will issue recommendations according to their use and successful redeployment. Appropriate dissemination measures will be (and, as you can see, are actually) taken by using Internet communication facilities.
Click here to return to the table of
contents
The applications' context
This section is divided into two subsections :
Click here to return to the table of contents
Motivations
The complexity of new (and under development) software intensive systems is increasing so much that new methods must be found. These methods should provide system engineers with rigor to assist them in their specification tasks, and also to help them guarantee that the inherent complexity of requirements will be satisfactorily tackled both at the development level and at the change level.
This trial application proposed by ALCATEL ESPACE is expected to experiment and to combine some leading edge techniques in order to help managing requirements changes in a more structured way. It is also a means to try to fulfill the so called `T0 + 12 months' action of ALCATEL ESPACE that aims at reducing the development cycles.
The Application Domain
ALCATEL ESPACE is involved in the S 2 programme as main contractor. This programme aims to deliver a system based on satellites to a national administration. The programme is decomposed into 3 systems: the satellites, the ground stations and a network system. The network system is a software intensive system and is estimated to more than 1 000 000 of source code lines.
The project was launched in 1990 and should be achieved by the end of 1996. The development team is of about 200 persons. The exploitation duration is specified for 20 years.
The associated project plan, presented in Appendix 1, distinguishes between three versions of the system: V1, V2.1 and V2.F (a fourth version, V3, is planned). Each of them is in fact a sensible incremental evolution of the previous one. With respect to the 2RARE project, it should be noted that:
The trial application will be limited to the so-called `plan management' subsystem that is part of the network system. This subsystem provides the establishment of communications between the ground stations. It looks like a switching system.
This subsystem is decomposed into 2 equipments in the V1 and V2.1 versions:
Today, more than 20 major change requests have been applied on this subsystem and several are in progress. These changes are handled by a team of about 4 permanent persons. Among these changes, the introduction of a new equipment is studied in the V2.F version: the Requests Arbitrating Equipment (RAE). This equipment aims at managing the transmission resources.
The requirements document of the subsystem contains 130 pages. The requirements documents of the 3 equipments contain respectively 230, 209 and 127 pages.
People participating in the requirements expression are system engineers with a solid background in their technical area (telecommunications), but they don't necessarily have skills in computerized systems development.
The requirements are established from a standard document which describes what should be given. Accompanying guidelines are available. For one year requirements have been numbered in the documents. There is no specific formal or semi-formal method used for the requirements modelling.
During the development, many changes may occur at any system definition level, most of them being due to the evolution of users' needs. Each change request is evaluated at all stages of the project to determine the technical, schedule and cost impacts. As currently the requirements (and the links between them) are not formalized, the impact analysis requires a rather important effort to achieve the consistency of a such complex and critical system.
The scope of the Trial Application
The Trial Application will be concerned with the `plan management' subsystem refered above. The V1, V2.1 and V2.F versions of the Management Plan subsystem will be modelled across the trial application. V1 and V2.1 modelling will permit to refine the method before to apply it in real time on the V2.F version.
A specific team is in charge of the 2RARE project in order to avoid the introduction of changes in the existing Network System project team. This team is composed of modelling specialists and a Network System project member (who is in charge of providing the Network System knowledge, providing the inputs to the trial application, validating the outputs and introducing the leading edge technologies inside the Network System change request treatment).
The Trial Application team doesn't depend directly to the Network System project leader. Nevertheless, the trial application is performed in the same premises as the project Network System and a member of the Network System project will directly participate in the 2RARE project. During the V2.F processing task, the occurring change requests will be treated in real time by the trial application team and the outputs of the impact analyses realized in the trial application will be used by the Network System team.
In order to maintain several versions of the system consistent and to keep system requirement traceability available, the purpose of the Trial Application is to investigate and evaluate the use of technologies that offer:
Metrics will be calculated during the trial application in order to estimate the benefits obtained by applying the leading-edge technologies. These metrics are defined in Sect. 3.1.1.
Impacts on the Network System project will happen through:
The specifications produced for the chosen trial application are confidential in industry. But the impact of the trial applications expressed in terms of the project plan will be made public.
With respect to the marginal costs inherent to trial of the new leading-edge technologies, it is expected that they will arise from the following additional tasks for the project team:
These tasks represent an overload of 10% for the Plan Management team.
Click here to return to the upper level
Motivations
BELGACOM has the following requirements engineering needs:
The Application Domain
One of the most appealing applications of the ``information superhighways'' is the access - through a network - of multimedia information. Applications like video-on-demand, teleshopping, tele-education, etc. will become possible in the near future, thanks to new broadband switching and transport technologies like ATM, SDH and access technology (PON,ADSL,HFC,etc.). In all of these applications, a user at work or at home accesses a remote multimedia server and retrieves multimedia information from that server. This communication is typically asymmetric in nature: the communication bit rate from the user to the server is typically in the range of 1 to 20 kbit/s, and far less than the bit rate of the communication from the server to the user, which is typically a compressed video stream of e.g. 2 Mbit/s. Because of this asymmetric bit rate , applications like video-on-demand and teleshopping are called retrieval services, in contrast with applications like multi-party video conferencing with a more symmetric kind of communication, and a more demanding time responsiveness (bidirectional video).
The new Multi Media division of BELGACOM will start a prospective project in the field of Broadband MultiMedia Retrieval Services (BMRS), aiming at the realization of business plans for different scenarios in the deployment of BMRS services.
The timing of the BMRS project is structured in two phases, each phase producing a technico-economical report of BELGACOM's MultiMedia market (alpha and beta versions of business plans). The associated project time schedule is produced in the figure presented in Appendix 2, the total manpower effort is about 49 man/month (spread over 1995 and 1996).
On the one hand, the BMRS project encompasses a technology workpackage that incorporates the TECTRIS technical (non-commercial) trial on a Video-On-Demand (VOD) service, and the study of the functionalities and the implementation aspects of a service gateway. This Service Gateway will provide the user with the possibility of choosing between different servers (possibly from different service providers) in a transparent way; according to its selection , the user is connected to the right server. The gateway plays a major role in the billing aspects of such a distributed service. This first technical trial with the video-on-demand (VOD) application has been scheduled for October 1995 (see workplan in the Appendix 2). The aim of the trial is a pure technological one, i.e. it is supposed to evaluate transmission technology (ATM) and the hardware platforms against the VOD like service. The specifications for this trial approved in January 1995 were essentially resource driven (bit rates, time and space efficiency). No real effort has been spent in the definition of specific VOD services (functionalities from a user and operator point of view).
On the other hand and at the same time , the BMRS project defines a market study workpackage to determine the user requirements for specific applications (e.g. teleshopping, on-line commerce, other interactive information retrieval) to be further implemented on the technical platform evaluated in the technology workpackage. By March 1996, this will result in a deployment plan for the introduction of these services.
The scope of the Trial Application
BELGACOM proposes to participate in the ESPRIT 2RARE project in three phases: the first two phases (up to beginning of March 96) will evolve along the BMRS technology workpackage, the last phase will depend on the MultiMedia plans produced at the outset of the BMRS project:
The ESPRIT 2RARE team will cooperate with the application specification team: the former using leading-edge requirements engineering technologies brings a complementary approach to BELGACOM specification work of the latter. The added value of 2RARE is to bring techniques allowing for rapid changes in requirements, for additions of new services and features ...
Consequently a marginal effort of 30% is expected: the learning effort in front of novel technologies (tools) and concepts (formal language semantics) accounting for 10% additional efforts; the expression (formulation) of formal representations in full detail depending on the expressive power of the modelling languages accounting for 20% of additional efforts. However, if the work in ESPRIT 2RARE is successful, a high re-utilization rate of the spent effort is expected: the requirements engineering expertise will be reused on similar multimedia service specification projects and be based on the management of changes and their replay applied on previously produced specifications.
Click here to return to the upper level
Click here to return to the table of
contents
The underlying technologies
Since one main goal of the project is to evaluate the benefit of using and
integrating leading edge RE techniques for tackling and achieving a better
understanding of the requirements management problem, we describe here
which are those techniques according to their purpose.
Click here to return to the table of
contents
Project snapshots
Click here to return to the table of contents
Click here to return to the upper level
PROJECT N º : | 20.424 | ACRONYM : | 2RARE |
|---|---|---|---|
Target Markets:Mention main market sectors in which the partners (both users and developers) plan to commercialise or benefit from results of the project. | |||
|
Telecommunication: software-based development, development of intelligent
network services, Satellite communication systems (civilian and
military), Ground control and monitoring system, satellite
network systems. Aeronautics and space: development of embedded safety critical software intensive systems. CASE tool development for requirements engineering. Education, Community of requirements engineering researchers. | |||
Technologies:Mention the key technologies developed/used in the project. | |||
| The key technologies used are (i) formal techniques for requirements specification (Albert and Oblog) as well as their supporting tools environment and (ii) the Doors traceability tool. | |||
Exploitable results:Describe the results that are planned to be exploited, and in which form (services, products, improved processes), in relation to identified user needs. | |||
|
ALCATEL: Methods and tools for requirements management and system modeling. BELGACOM: Improved processes for service creation in the context of Intelligent Network Services. Development of an adequate methodology for dealing with changing requirements for such service. IXI: The main exploitable result that emerge from 2RARE is at the moment a demonstration of our skills in complex system object-oriented modeling. This corresponds to services that we deliver, but this is important for us in the context of a large project such as ALCATEL's. The global requirements management approach that is under experimentation within 2RARE could also be very interesting to promote as a global methodological support. It is however a bit early to conclude about this point. We prefer to wait for the end of 2RARE and to get some consolidated feed-back before starting exploitation actions around this topic. OBLOG: The users in the 2RARE project gave different uses to the OBLOG CASE. Both usages revealed weaknesses and capabilities of OBLOG. We think the limitations that, up to now, BELGACOM and ALCATEL pointed out, are due to limitations of OBLOG CASE current version and can not be imputed to characteristics of the specification language. The capabilities showed in the trial applications are a positive reassessment, indicating that OBLOG has a potential to be successful in applications in the technical area. ORIGIN: experiment with OO methodology combined with a structured approach during the whole project cycle even in critical environment (unstable users requirements, quick evolving specifications, high technical environment, ...), leading to a definition of an improved development methodology CERT: is aware of the main role played by traceability when validating some formal designs. Its knowledge in the field of formal methods is improved a lot by the use of such a classical way to check the correctness of industrial specifications and designs. FUNDP: Feedback on the Albert language and on the tools required for supporting its use VUA: Exploitation of the results achieved in the area of traceability of requirements as well as in the area of requirements structuring. | |||
Anticipated and demonstrated benefits:Describe the measurable benefits that each of the exploitable results is expected to provide or demonstrates to the target markets. | |||
|
ALCATEL: The benefits could be measured only by
demonstrating the ability to deliver the right systems in
conformance with the schedule. The benefits could be also
measured through the number of changes applied during the system
development: by applying riguorous techniques, the problems should
be identified earlier and the requirements become quickly stable. BELGACOM: shorter time-to-market time for the development of a new service. IXI: Improving requirements management is a necessity in most of spatial projects in which times of delivery are in constant diminution. The use of object-oriented modeling techniques enables a better efficiency (whose quantification is outside the scope of 2RARE) in the sense that some specification mistakes can be avoided early in the project. Those mistakes have not to be dealt with in the later stages of a project. A global requirements management approach can go further in the sense that some problems as change management can be addressed with a well defined process. OBLOG: Some requirements were common to both users. These will be evaluated and progressively incorporated in the next versions of OBLOG CASE, in order to meet anticipated needs of the technical market. ORIGIN: Expertise in critical environment, with a proven approach. CERT: the main demonstrated benefit is the set up of a complete and consistent System Design Life Cycle from Informal requirements to Formal designs. FUNDP: The use of Albert as an expressive formal specification language in terms of which requirements expressed by customers can be represented VUA: Improvements of our education and of our research products TCM and TRADE. and requirements structuring | |||
Application and assessment of results:Describe the specific applications where the technologies are deployed in the context of an action, as well as any outcome from thei assessment | |||
|
ALCATEL: as far as we will be able to measure the effectiveness of the
methodology, we plan to rely on it for the development of new
software intensive systems. BELGACOM: use of the technologies in the context of the development of a calling card system IXI: The requirements management approach is used for subsystem GdC of the S2 system (French military telecommunications systems). - see ALCATEL Forms -. OBLOG: working prototype for integration of Oblog with Doors. Improvement of documentation generation capabilities ORIGIN: Not yet. As soon as the results will be known, presentation of an improved development toolkit CERT: Usability of Formal Methods for Safety Critical Software Intensive System. FUNDP: On the basis of the feedback given by users, we are working on the improvement of some Albert language constructs and revisiting some semantic issues. VUA: We will use the results of 2RARE in our education because we can give examples from the 2 case studies of 2RARE. Of course, we will respect confidentiality requirements of the participating companies. We also use the results in the definition of the functionality of a requirements engineering workbench that we are currently producing, called TCM (Toolkit for Conceptual Modeling). (Because we are an academic institution, TCM is public domain software, available through anonymous ftp.) TCM supports the TRADE toolkit. TRADE (Toolkit for Requirements and Design Engineering) is a collection of conceptual tools techniques and heuristics; TCM is a collection of graphical editors by which the TRADE techniques can be used. 2RARE is particularly useful in specifying a conceptual model for TCM that includes the required traceability links. It also is useful in verifying our ideas about requirements structuring. | |||
Innovative aspects:Describe the innovations achieved with respect to the current state of the art. | |||
| The innovation of 2RARE is the combination of formal and semiformal requirements specification techniques in the context of two large industrial projects. Moreover, another innovation comes from the elaboration of a methodoly for eliciting the requirements at different levels of abstraction and for keeping them consistent across the changes and the releases managed in parallel. | |||
Click here to return to the upper level
PROJECT N º : | 20.424 | ACRONYM : | 2RARE |
|---|---|---|---|
Target Markets:Mention main market sectors in which the partners (both users and developers) plan to commercialise or benefit from results of the project. | |||
|
Telecommunication: software-based development, development of intelligent
network services, Satellite communication systems (civilian and
military), Ground control and monitoring system, satellite
network systems. Aeronautics and space: development of embedded safety critical software intensive systems. C.I.M. and industrial plants control. CASE tool development for requirements engineering. Education, Community of requirements engineering researchers. | |||
Technologies:Mention the key technologies developed/used in the project. | |||
|
Used leading-edge technologies are (i) formal specification techniques for requirements specification
(ALBERT and OBLOG) as well as their supporting tools environment and
(ii) the DOORS traceability tool. During the project, the use of these leading-edge technologies was complemented with the use of other RE techniques like the PSS-05-0 ESA standard template for RE document and the ObjectGEODE CASE tool | |||
Exploitable results:Describe the results that are planned to be exploited, and in which form (services, products, improved processes), in relation to identified user needs. | |||
| Within the context of two large and complex distributed software-intensive heterogeneous systems, the objective of the project was to use the leading-edge technologies (see above) with the obejctives of improving the management of changing requirements through the production of RE documents of a better quality (structuring, absence of ambiguities, inconsistencies, ...) | |||
Anticipated and demonstrated benefits:Describe the measurable benefits that each of the exploitable results is expected to provide or demonstrates to the target markets. | |||
|
The project made clear that the way according to which different RE techniques
can be used with the highest benefit was heavily depending on the dgree of maturity
of (i) the RE processes in place in the organization and of (ii) the used techniques. Since the RE maturity level was different for the two users (BELGACOM and ALCATEL ESPACE), the use and the benefits of the RE techniques appeared different. However, it is clear that the introduction of RE formal methods is only possible in organizations having solid experience in the production of high quality RE textual documents and conceptual modelling. | |||
Application and assessment of results:Describe the specific applications where the technologies are deployed in the context of an action, as well as any outcome from their assessment | |||
The different technologies (see above) have been used and evaluated in the following context:
| |||
Innovative aspects:Describe the innovations achieved with respect to the current state of the art. | |||
The project has drawn some lessons and recommendations on the following questions:
| |||
Click here to return to the upper level
Click here to return to the table of
contents
Public deliverables of the project
Deliverable D12 : Final Report on Users' Results(Word Document)
Deliverable D13 : Final Report on Suppliers' Results(Word Document)
Click here to return to the table of
contents
The people working on the project
Herve Joumier
E-mail : frdk82w8@ibmmail.com
Christian Le Mene
E-mail : frdk8px8@ibmmail.com
Pascal Very
E-mail : frdk8rj8@ibmmail.com
Francis Depuydt
E-mail : Francis.Depuydt@is.belgacom.be
Peter Dieu
E-mail : Peter.Dieu@is.belgacom.be
Luk Levrouw
E-mail : Luk.Levrouw@is.belgacom.be
Jack Foisseau
E-mail : Jack.Foisseau@cert.fr
Michel Lemoine
E-mail : Michel.Lemoine@cert.fr
E-mail : shuyts@ctp.com
E-mail : roelw@cs.vu.nl
Marc Vinesse
Jean Ribes
Philippe Herbin
Jean-Noel Mathon
Joao Gouveia
E-mail : jgouveia@oblog.pt
Pedro Inacio
E-mail : pinacio@oblog.pt
E-mail : edu@info.fundp.ac.be
E-mail : phe@info.fundp.ac.be
Click here to return to the table of contents
Click here to return to the Computer Science Dept. home page.
Patrick Heymans -
November 13, 1995Last update : October 22, 1996