Towards an Understanding of Performance, Reliability, and Security

Introduction

According to J. Spohrer, service-oriented systems are “value co-creation configurations of people, technology, internal and external service systems connected by value propositions, and shared information (such as language, laws, measures, models, and so on)” (Spohrer, 2007). A service-oriented system (Espadas et al., 2013) delivers a set of services that are cooperating with each other over the Internet, which are therefore called “web services”. Web services are delivered in a new computing paradigm called “service-oriented computing”. Service-oriented computing paradigm plays an increasingly crucial role in modern world industries. Even in cloud computing, services are irreplaceable elements (Wang & Sun, 2015). In service-oriented systems, software developers create software applications through composing web services. Web service composition helps developers to solve complex problems by combining published basic services and ordering them to meet specific service requirements from user requests (Song & Lee, 2013).

Service-oriented systems offer many benefits to software developers and software users. For example, they allow developers to support many customers with a single version of software and provide services to their customers 24/7. Over the past decade, a large number of service-oriented systems have been developed to aid a wide range of real-world applications (Wang et al., 2011). For example, E-tourism service systems (see for example: www.paypal.co.uk), an online payment service, has been used by many e-commerce applications worldwide. Many investment banks or brokerages nowadays offer their customers trading services such as order management services, stock trading services, bond trading services and so on, and have been playing an increasingly pivotal role in the financial sector (Yang et al., 2011).

There are a lot of key research issues with regard to service-oriented systems, involving service modeling and specification (Cardoso et al., 2010; De Virgilio, 2010), service selection (Sweeney et al., 2016), web service composition algorithms (Hiel & Weigand, 2010; Leitner et al., 2010), verification of composition correctness (Rai et al., 2015), and handling composition faults (Lemos et al., 2016), etc. Rather than only concentrating on functionality, recent work puts more emphasis on Quality of Services (QoS) properties (Calheiros et al., 2015). With many businesses providing same or similar services, QoS properties have become an important indicator for good services (Zhao et al., 2007). Yet, satisfying QoS requirements in service-oriented systems has proved to be more challenging than satisfying functional requirements, due to the following characteristics (Yang et al., 2011): First, QoS properties are system-level requirements which cannot be assigned directly to individual system components; instead, they need to be planned at the infrastructure-level as a whole, with their design aspects then entrusted to system components. Second, QoS properties are often interdependent and their realization in a system requires a collective and coordinated behavior of the system components and a system-level design strategy. Third, QoS properties are application-specific and their fulfillment necessarily requires a specific design approach germane to their applications. Finally, QoS properties are not only a design time concern, but also most crucially a runtime concern. Satisfying QoS properties means designing runtime mechanisms that can maintain the system’s QoS properties throughout the execution time.