Comprehensive approach to service interaction handling1

https://doi.org/10.1016/S0169-7552(98)00149-4Get rights and content

Abstract

To achieve effective and efficient handling of service interactions in an industrial setting, a comprehensive approach is needed. The EURESCOM Project P509, “Handling Service Interactions in the Service Life Cycle”, proposes such an approach providing both technical and organisational solutions to the problem. This novel approach treats interaction handling as a part of service integration rather than service creation. The paper presents the most important elements of the approach: the concept of Interaction Filtering increasing the efficiency of pre-deployment interaction detection; the Service Interaction Handling Process enabling interaction detection on a large scale and providing a framework for interaction solving; new enterprise roles responsible for interaction handling; and a set of off-line and on-line interaction handling mechanisms used by the approach. The evaluation of the mechanisms based on the results of an extensive case study is presented as well.

Introduction

Deregulation of telecom markets and increasing customer demands create new challenges for service providers. They have to promptly react on changing market requirements and provide new services more rapidly than in the past. Though advanced service architectures such as the Intelligent Network (IN) or Telecommunications Information Networking Architecture (TINA) facilitate rapid service provisioning, there are several obstacles in achieving this goal. The service interaction problem is commonly regarded as one of them.

We usually talk about service interaction when the behaviour of one service is affected by the behaviour of another service or another instance of the same service. Interactions can be the reasons of various problems, such as user dissatisfaction, misuse of network resources, deadlocks, etc. They occur mainly because new services and features exceed the original design limitations of the underlying networks, and because new services are developed independently upon different, often inconsistent assumptions. There can be, of course, many other causes of this phenomenon.

Undesired interactions deteriorate the quality of services. Therefore, telecom industry must deal with the interaction problem in one way or another to meet its objectives of speed and quality in the provisioning of new services. Considering the fact that the number of potential interactions grows exponentially with the number of services and their features deployed in a network, finding efficient ways to handle the interaction problem becomes imperative.

Since interactions occur in many forms and can be introduced during nearly every phase of the Service Life Cycle (SLC), interaction handling requires multiple, complementary mechanisms to cover the whole problem. In addition to off-line and on-line mechanisms for interaction detection and resolution, organisational solutions are necessary as well. Interactions do not only have impact on systems, but also on business processes and functions such as customer care, billing, and service management. Therefore, in order to properly handle the interaction problem on an industrial scale a comprehensive approach is needed.

In this paper, we present a proposal of such a comprehensive approach developed by the EURESCOM Project P509 “Handling Service Interactions in the Service Life Cycle”. The approach includes both technical and organisational solutions that are packaged into a new telecom business process. The concept of a new process separates the task of interaction handling form service creation and treats it as a part of service integration, instead.

From the very beginning, EURESCOM has recognised the importance of the interaction problem for the European Public Network Operators4 (PNOs). A number of EURESCOM projects such as P103 “Evolution of IN”, P230 “Pan-European IN”, P509 and others addressed different aspects of the service interaction problem.

The EURESCOM Project P509 “Handling Service Interaction in the Service Life Cycle” was entirely devised to the issue of service interaction. The project started in 1995 and ran for two years. P509 was a co-operative work of eight European PNOs: Tele Danmark, Finnet Group, France Telecom, Deutsche Telekom, KPN, OTE, Telefonica, and Telia. P509 was probably the largest international project in the area of service interaction so far.

Results of project P509 including the comprehensive approach to interaction handling have been previously published in EURESCOM internal reports (see 13, 14) as well as in a number of conference papers (see 2, 5, 6, 7, 8, 9, 10, 11, 12). In this paper, we would like to summarise the project's results and highlight the most important elements of the P509 approach.

In the next section, we summarise the state of the art in the field of interaction handling and motivate the directions taken by P509. In the subsequent sections, we present the most important contributions of the P509 approach, such as, the idea of Interaction Filtering; the concept of Service Interaction Handling Process; as well as two new enterprise roles of Service Integrator and Service Mediator responsible for interaction handling. Furthermore, the off-line and on-line mechanisms used by the P509 approach are outlined and evaluated.

Section snippets

Rationale behind the P509 approach

It is our view that the current industrial practice in the field of service interaction handling is still insufficiently mature. This does not mean, of course, that nothing is done in this matter. Telecom manufacturers (mainly the vendors of switching systems), who have been dealing with the interaction problem for over 20 years, certainly have developed internal procedures and quality standards to deal with the undesired interactions between the services and features they create and provide in

Summary of the P509 approach to service interaction handling

The main goal of EURESCOM Project P509 was to provide European PNOs with a comprehensive approach to interaction handling and with useful guidelines for addressing the interaction problem in their organisations.5 In particular, the approach was to facilitate effective and efficient application of interaction detection methods on an industrial scale.

The scope of the

Separation of interaction handling from service creation

For historical reasons, PNOs have been playing multiple roles, such as Network Operator, Service Creator, and Service Provider. Likewise, suppliers of switching systems have been playing roles of Platform Vendors and Service Creators. This multiple-role character of telecom organisations has strongly influenced the idea on where and how service interaction handling should be done.

Interaction detection and resolution is often seen as an integral part of service creation [3], which seems to be

Interaction handling mechanisms

During the initial state-of-the-art study, the various methods and mechanisms proposed in the literature were assessed and several of them were selected for adoption (and sometimes adaptation) by the P509 project. Thereby, the most important selection criteria were the potential of methods and mechanisms for application in industrial practice, i.e. efficiency and scalability. An objective was to choose a set of mechanisms that complement each other rather than overlap. Therefore, the mechanisms

Evaluation of the interaction handling mechanisms

The applicability of the mechanisms described in the previous section was evaluated during an extensive case study. Although this study did not involve actually implemented services or field trials, it provided an assessment of the usefulness of the various mechanisms, both in their own right and as part of a comprehensive approach.

Below we present the scope of the P509 case study and summarise its results. Each type of mechanisms is briefly evaluated in a separate sub-section. Additionally a

Summary and conclusions

The interaction problem in telecom systems has many dimensions and can occur on different levels of abstraction and in different phases of the Service Life Cycle. Undesired interactions can deteriorate the quality of services and cause dissatisfaction of their users and subscribers. Since the number of potential interactions grows exponentially with the number of services and features deployed in a network, the efficient handling of the problem might be very time consuming and may disable rapid

Kristofer Kimbler received his M.Sc. in Computer Science from the University of Warsaw in 1983. In 1984, he joined the Warsaw Institute of Technology and four years worked with the development of computer tools for SDL and CHILL. In 1989 he moved to Sweden, and till 1992 worked as a consultant with advanced PABX applications. In 1993, he started his Ph.D. program at Lund University, and in 1996, received the Technical Licentiate degree in the area of service interaction. From 1995 till 1997, on

References (15)

  • J. Rumbaugh et al., Object-Oriented Modelling and Design, Prentice-Hall, Englewood Cliffs, NJ,...
  • I. Aggoun, P. Combes, Observers in service creation and service execution, in: Feature Interactions in...
  • R.B. Blumenthal, M.P. O'Reilly, P. Russo, Addressing feature interactions during service creation, in: Feature...
  • C. Capellmann, R. Demant, R. Galvez-Estrada, U. Nitsche, P. Ochsenschläger, Case study: service interaction detection...
  • P. Combes, B. Renard, C. Capellmann, J. Pettersson, J. Ruiz, Consistent and comprehensive approach to off-line service...
  • P. Georgatsos, T. Nauta, H. Velthuijsen, Role of service management in service interaction handling, in: Feature...
  • K. Kimbler, Addressing the interaction problem at the enterprise level, in: Feature Interactions in Telecommunication...
There are more references available in the full text version of this article.

Cited by (3)

  • Composing features and resolving interactions

    2000, Proceedings of the ACM SIGSOFT Symposium on the Foundations of Software Engineering

  1. Download : Download full-size image
Kristofer Kimbler received his M.Sc. in Computer Science from the University of Warsaw in 1983. In 1984, he joined the Warsaw Institute of Technology and four years worked with the development of computer tools for SDL and CHILL. In 1989 he moved to Sweden, and till 1992 worked as a consultant with advanced PABX applications. In 1993, he started his Ph.D. program at Lund University, and in 1996, received the Technical Licentiate degree in the area of service interaction. From 1995 till 1997, on behalf of Telia Research, he worked as the leader of EURESCOM project P509 “Handling Service interaction in the Service Life Cycle”. Currently, Kristofer is involved in the ACTS project TOSCA “TINA Open Service Creation Architecture” and completes his Ph.D. thesis. He also works for High Definition Systems AB as a consultant in the area of software engineering and telecommunications.

  1. Download : Download full-size image
Carla Capellmann studied Computer Science at the University of Koblenz. After having graduated in 1991, she joined Deutsche Telekom AG, Technologiezentrum at Darmstadt. She is member of research and development department “Platforms for value-added services, network and service management” where she currently leads a project in the area of IN and TINA. Before, Carla was involved in several EURESCOM projects in the area of IN. In 1995 and 1996, she was a task leader EURESCOM Project P509.

  1. Download : Download full-size image
Hugo Velthuijsen received his M.Sc. from the University of Amsterdam in 1986. Upon graduation he joined KPN Research (then called Dr Neher Laboratories), the R&D organisation of PTT Telecom. His initial interests were mainly concerned with Artificial Intelligence, leading to a Ph.D. in 1992 from the University of Maastricht. Also in 1992, Hugo spent a year as visiting researcher at Bellcore, where he was introduced to the feature-interaction problem. He was Programme Chair of the First Feature Interaction Workshop in St Petersburg, FL in 1992 and Workshop Co-Chair for the Second Feature Interaction Workshop in Amsterdam in 1994. In 1995 and 1996, Hugo was involved in EURESCOM project P509 as a task leader. Since June 1997, Hugo Velthuijsen works as business consultant at PTT Telecom.

1

This paper is an extended version of the paper “Addressing the interaction problem at the enterprise level” presented on the 4th Feature Interaction Workshop in Montreal, June 1997 (see [7]).

2

E-mail: [email protected].

3

E-mail: [email protected].

View full text