A system for monitoring, assessing and certifying Quality of Service in telematic services

https://doi.org/10.1016/j.knosys.2007.02.003Get rights and content

Abstract

This work presents a knowledge-based system and architecture for the formalization of quality of service (QoS) characteristics and measurement methods in telematic services, and for the collection, distribution and assessment of the QoS information obtained using these methods. The solution makes use of ontologies and reasoning technology to define and compare QoS information. The system can help customers to select the service offer that best fits their quality requirements. Providers will use it to detect when the delivered service fails to fulfill the quality agreements that they signed with customers.

Introduction

Telematic services are those provided through a telecommunication network and involving information exchange (for example: data, video or audio). As new services appear with high performance requirements, mechanisms to ensure quality of service and metrics to monitor this quality become necessary.

The right selection of a service provider by the customers depends on the availability of information about the QoS offered and delivered by each provider, so they may perform a comparison and make a decision. From the provider’s point of view, monitoring QoS is important to check the fulfillment of agreements. Service level agreements (SLAs) are the contracts that customers and providers sign to define the requirements of the services being delivered. Measurements and requirements must be compared in order to detect SLA violations. The providers may also want to use quality to differentiate their offer from those of the competitors. To do so, they would make QoS data available to quality agencies or other independent actors to let them certify that the provider delivers a particular level of quality.

The methodology, mechanisms and characteristics used to measure the QoS should be the same for all the providers, so the measured data can be used for comparisons. This is not always possible due to the provider’s business strategies. A possible approach is making independent actors, such as quality agencies or organizations for standardization responsible for defining characteristics, metrics and reference service levels (RSLs) and checking if the measurements taken at the provider’s infrastructure fulfill the defined requirements.

Two more tasks are required: the design of a method to publish and exchange the QoS information and the implementation of a system to compare the measurements against requirements from SLAs or RSLs. Knowledge-based technology, such as the Semantic Web [1] approach, contributes to simplify the development of these systems. The Semantic Web defines a framework for data sharing and information reusability. The formalization of the data by means of ontologies [17] facilitates sharing and exchanging information, also reducing the possibility of misunderstanding the meaning of the concepts. Rules can help to automate the comparisons by defining how instances of the ontologies can be handled to perform the evaluation. So, the Semantic Web characteristics of data sharing, distribution and rule reasoning are useful for the environment where telematic services live.

This paper presents a knowledge-based system for the definition and formalization of QoS characteristics and measurement methods, and for the collection, distribution and assessment of the QoS information obtained using these methods. First, the system architecture to exchange QoS data is presented and the responsibility of each participant is identified. Then, the ontologies used in the system are defined. Later, the rules driving the system decisions about quality are described. Finally, a test scenario with Internet Service Providers (ISP) is used to demonstrate the functionality of the system.

Section snippets

System architecture

The system presented in this paper develops a formal definition of all the terms involved in the quality of service computation. As stated before, it uses ontologies for this purpose. It also introduces a methodology based on ontology instances to exchange information about measurements and QoS requirements (from RSLs and SLAs). The designed system avoids problems derived from misunderstanding the meaning of that information by providers and independent actors thanks to the formal definition of

QoS ontologies

A system composed by applications designed by different companies that want to share information could be complex due to interoperability problems. The definition of models using ontologies allows a formalization of concepts and makes easy to understand the information required for a system to work. Consequently, it reduces complexity when programming an application or exchanging data.

One of the design goals required in ontologies is that actors agree to share the ontology. So it is necessary

Rules to define evaluation of QoS

Processing individuals to get the evaluation of the quality provided in a service offer could be done developing a specific application. This was the first approach we took, but some problems arose, such as the high workload to obtain acceptable results or the difficulties in producing a system abstract enough to work with any service, metric and characteristic.

A different approach uses rules to define the way of evaluating quality of service against QoS profiles. The existence of reasoners

Test scenario

To demonstrate the functionality of the system, a test was performed in a scenario with three providers that publish measurements for an ISP service based on ADSL connections (Table 1). An organization for standardization generates instances of the QoS ontology with the metrics and characteristics for this service (Fig 6). ETSI Metrics for the ISP service have been used [4]. A quality agency publishes a RSL for the ISP service including requirements with values for different metrics. In the

Comparison with other works

Architectures for QoS assessment are usually based on proprietary technologies or customized implementations. However, there are projects that have tried to provide more open and flexible solutions by modelling QoS using ontologies [3], [14], [19] and by defining knowledge based brokers to select the best service [14], [19]. These works are oriented to telecommunication services, some of them are specific to Web Services.

Regarding QoS modelling, none of the existing efforts provide an ontology

Conclusions

In this paper we have presented work carried out to cope with the current need of having a knowledge base and a system to perform QoS comparisons. This system can certify if a provider delivers a particular level of quality. This would be done comparing QoS measures against a RSL. It can also detect SLA violations. The comparison will be made between QoS measurements and an agreement in a SLA. Another function is the discovery of services with a delivered quality that matches the quality

Acknowledgment

This work has been partially supported by the Spanish National Plan of Research, Development and Innovation (Ministry of Education and Science) under Grants TIC2003-04406 and TSI2005-07306-C02-01.

References (20)

  • T. Berners-Lee et al.

    The Semantic Web

    Scientific American

    (2001)
  • Cisco. Cisco IOS NetFlow....
  • G. Dobson, R. Lock, I. Sommerville.QoSOnt: a QoS Ontology for Service-Centric System, in: Proceedings of the 31st...
  • ETSI. Final Draft ETSI EG 202 057-4. V1.1.1. Speech processing, transmission and quality aspects (STQ). User related...
  • I. Horrocks et al.

    OWL rules: a proposal and prototype implementation

    Journal of Web Semantics

    (2005)
  • I. Horrocks, P.F. Patel-Schneider, H. Boley, S. Tabet, B. Grosof, M. Dean. SWRL: A Semantic Web Rule Language Combining...
  • IST-SCAMPI. A Scaleable Monitoring Platform for the Internet....
  • IST-INTERMON, Advanced architecture for INTER-domain quality of service MONitoring, modelling and visualisation....
  • ITU-T. Recommendation E.430—quality of service framework, June...
  • ITU-T. Recommendation E.800—terms and definitions related to quality of service and network performance including...
There are more references available in the full text version of this article.

Cited by (4)

  • Introduction to quality of service

    2010, Intelligent Quality of Service Technologies and Network Management: Models for Enhancing Communication
  • Scenario-based configuration management for flexible experimentation infrastructures

    2009, 2009 5th International Conference on Testbeds and Research Infrastructures for the Development of Networks and Communities and Workshops, TridentCom 2009
View full text