Evolutionary Architecting of Embedded and Enterprise Software and Systems

Introduction

In many companies developing technical products, such as the automotive industry, process automation, defense, or telecommunication, embedded systems and software play an increasingly important role. The embedded systems have developed from simple, stand-alone processors into a large number of computers with distribution networks and millions of lines of software. Each processor node can have many actuators and sensors, which are sometimes very information intensive. There are also often links (wired or wireless) to other systems for information exchange. The embedded system needs to be packaged in whatever space is available in the product, and be equipped with an adequate power supply system, which often introduces both new operating modes, and electrical requirements that have implication on the software in different ways. (For an overview of contemporary complex embedded systems and their challenges, see e.g. Broy et al., 2007; Grimm 2003).

This increasing complexity leads to soaring development costs, and many companies strive to curb this trend by reusing software and hardware between products. At the same time, the development cost targets have to be balanced against the strict requirements of these products. They are often business-critical, or even safety-critical; they are installed in harsh environments and must be very robust against disturbances; they have a long lifetime where it is usually not acceptable to discard the overall product just because of a failure in the embedded system; and finally, they often have to be flexible to adapt to changes and upgrades after delivery to customers. This means that a substantial part of the design effort deals with supporting other lifecycle stages than just normal operation.

With these requirements, and a multiplicity of products and variants for different customers and markets, the architecture is becoming very important and is a source of increasing interest for companies developing embedded systems. An architecture can be defined as “the fundamental organization of a system embodied in its components, their relationships to each other, and to the environment, and the principles guiding its design and evolution” (IEEE, 2000). Often, a product line approach is applied, where the same platform is used as a basis, with modifications to fit individual products and customers. The decisions made by architects in the early phases influence many decisions made later on, and the architecting decisions are difficult to change further down the process (Smith and Reinertsen, 1995). With a poor architecture, downstream development activities will thus become much more expensive and time consuming.

In the embedded systems literature, there is a strong focus on systems and hardware architecture (see e.g. Sangiovanni-Vincentelli and Di Natale 2007), and this is natural since custom hardware is almost always used and there are hence many degrees of design freedom in the choice of hardware components and structure. Over the years, this focus has been complemented with increasing emphasis on software architecture as a consequence of the rising importance of software in these products. However, what is often neglected in the literature is a discussion on the links between the embedded systems in the product and the enterprise architecture, i.e. the organizing logic and IT infrastructure, of the firms that produce it or interact with it over its lifetime.