Abstract

Services composition is fundamental to software development in multi-service wireless sensor networks (WSNs). The quality of service (QoS) of services composition applications (SCAs) are confronted with severe challenges due to the open, dynamic, and complex natures of WSNs. Most previous research separated various QoS indices into different fields and studied them individually due to the computational complexity. This approach ignores the mutual influence between these QoS indices, and leads to a non-comprehensive and inaccurate analysis result. The universal generating function (UGF) shows the speediness and precision in QoS analysis. However, only one QoS index at a time can be analyzed by the classic UGF. In order to efficiently analyze the comprehensive QoS of SCAs, this paper proposes an improved UGF technique—vector universal generating function (VUGF)—which considers the relationship between multiple QoS indices, including security, and can simultaneously analyze multiple QoS indices. The numerical examples demonstrate that it can be used for the evaluation of the comprehensive QoS of SCAs subjected to the security constraint in WSNs. Therefore, it can be effectively applied to the optimal design of multi-service WSNs.

Highlights

  • In recent years, wireless sensor networks (WSNs) have evolved with new features such as large scale, high device heterogeneity and the capability of supporting multiple applications [1]

  • In order to assure the lowest security strength of services composition applications (SCAs) in WSNs satisfying users’ individualized security requirements and improve other quality of service (QoS) indices as much as possible, this paper studies the analysis method for the comprehensive QoS with security constraints of SCAs in WSNs based on the vector universal generating function (VUGF) technology

  • Our contributions are as follows: (I) We present the working mechanisms of SCAs in WSNs and analyze the necessity of comprehensive assessment on multiple QoS indices; (II) We analyze the disadvantages of the classic universal generating function (UGF), and present a VUGF technique that can obviously improve the efficiency on comprehensive QoS assessment in a parallel computing environment; (III) Employing our VUGF, we present a composition calculation method of multiple QoS indices for

Read more

Summary

Introduction

WSNs have evolved with new features such as large scale, high device heterogeneity and the capability of supporting multiple applications [1]. Some approaches have been presented for the seamless integration of WSNs with these SOA technologies to successfully build a service application based on wireless sensor networks [9,10,11]. Unlike the traditional software QoS assurance technique, the QoS assurance method for the application of services composition in WSNs pays more attention to the mechanism of flexible QoS measures, deduction and adoption based on cumulative evaluation of operation information in an open running environment [27]. Due to the open and dynamic running environment, continuously variable user’s requirement, randomly selected SNs and the own characteristics of loose coupling and long transaction, the QoS optimization for the application of services composition in WSNs are confronted with severe challenges, which seriously restrict the further development, application and extension of the SCAs [31].

Working Mechanisms of WSNs Service Broker
VUGF Technique
Classical UGF Technique
Vector UGF Technique
Definition of Composition Operator
Composition Calculation for the VUGF of Two Random Vectors
Recursive Computation of the VUGF of Multiple Random Vectors
Analysis of Multiple QoS Indices Including the Security
Calculating the VUGF of QoS Indices for Each Atom-Service
Calculating the VUGF of the QoS Indices for Each Composite Service
Calculating the VUGF of the QoS Indices for Entire SCA Instance
Estimating the Comprehensive QoS of an SCA Instance
Description for the Example
Calculating the VUGF of Each Atom-Service
Calculating the VUGF of Each Composite Service
Calculating the VUGF of the Entire SCA Instance
Findings
Conclusions and Future Work
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call