Article Preview
TopIntroduction
The terminology concerning integration and interoperability is not universally consistent, given its wide breadth of application, from simple data up to complex systems and their interaction behaviors. Throughout the literature, the two terms are sometimes used inconsistently, depending on context, or even interchangeably.
It starts by the inexistence of a universally accepted definition of interoperability, since its meaning can vary accordingly to the perspective, context or domain under consideration. Although limited to information, the ISO/IEC/IEEE 24765 standard (ISO, 2010) provides the seemingly most cited definition of interoperability, as “the ability of two or more systems or components to exchange information and to use the information that has been exchanged”.
Reducing the context to Enterprise Information Systems (EIS), the EN/ISO 19439 standard (ISO, 2006) refers to enterprise integration as the process of ensuring the interaction between enterprise entities necessary to achieve domain objectives.
Therefore, interoperability asserts the ability of two systems to understand each other’s messages, whereas integration requires collaboration to achieve common goals. Interoperability is thus necessary but not sufficient to achieve enterprise integration (Chen, Doumeingts, & Vernadat, 2008), which usually entails cooperation and coordination at higher abstraction levels. This paper concentrates on interoperability, as a means and an enabler to achieve enterprise integration.
Just ensuring interoperability is already a daunting task, given the complex, heterogeneous and highly variable enterprise networks that characterize today’s enterprise IT landscape, in particular when considering the most recent, game-changing developments, such as cloud computing (Mezgár, & Rauschecker, 2014), Big Data (Marz, & Warren, 2015; Reed, & Dongarra, 2015) and Internet of Things applications (Want, Schilit, & Jenson, 2015).
Figure 1 illustrates a scenario that may well be applicable to many enterprises, to interconnect either subsystems of an EIS or complete EIS, in value chains or other cooperation agreements. An EIS will have to deal with conventional web servers, in-house servers, several general-purpose clouds types (private, public and hybrid) and specific cloud-like systems, involving mobility (Fernando, Loke, & Rahayu, 2013), sensor networks (Rawat, Singh, Chaouchi, & Bonnin, 2014) and RFID tags (Aggarwal & Han, 2013), increasingly used for supply chain control. All this needs to be interoperable, so that enterprise integration, as whole, can be achieved.
Figure 1. An example of a complex enterprise IT environment, needing integration
Taking into account that enterprises need to cooperate and to be connected, and the existing diversity of platforms and applications, any computer-based system can be considered a distributed and heterogeneous EIS. Some will be more dynamic than others, but all will endure the same basic problem: how to provide enough interoperability to enable EIS to cooperate without requiring exact mutual knowledge and allowing them to evolve independently. In other words, the tradeoff is how to provide the minimum EIS coupling to ensure interoperability but no more than that, to avoid hampering EIS evolution due to unnecessary dependencies.
The main goals of this paper are:
- •
To contribute to the field of enterprise architecture in the interoperability arena, by analyzing several of the problems involved, especially in what concerns the design of EIS;
- •
To describe a multidimensional framework that explores EIS interoperability at various levels, to better dissect and understand what is involved and how this can help in systematizing methods and solutions;
- •
To assess the potential of structural compatibility (instead of schema sharing, used in current interoperability technologies) as a means to reduce coupling and to increase adaptability, changeability and reliability, while maintaining interoperability requirements.