5.1 Introduction

As mentioned in Chapter 2, today's computing architecture is built according to the OSI model, ^[1] which focuses on the hardware and supporting technologies rather than the user . In order to realize a true me-centric computing environment, machines must understand users, not the other way around.

^[1] http://www.cisco.com/univercd/cc/td/doc/cisintwk/ito_doc/introint.htm#xtocid5

Have a look at Figure 5.1. Here you can see how the future computing model should look. It centers around the user. Through smart human-computer interfaces, users will communicate with intelligent appliances that were built for a specific purpose. Depending on the machine's capabilities, a given task can be executed by the appliances autonomously or can pass on certain subtasks to Web services on the Internet. The connection between the Internet and the appliances is handled by the intelligent network that is the networking infrastructure for me-centric computing.

Figure 5.1. Future Computing Model

The basic idea of the me-centric computing architecture is easily summarized: First, make computers disappear into intelligent appliances that everyone finds easy to employ . Second, have the appliances take on human work and accept delegated responsibilities, rather than asking humans first to divide tasks into those parts that they will do themselves and those parts that they must steer the computers to do. Then coordinate and combine the results coming from the myriad parts.

The me-centric computing model holds that inexpensive, microscopic processors and radio transceivers are embedded in everyday things, thus making everything " invisibly " intelligent. Today, we already find them in phones, television sets, cameras , cars , thermostats, and credit cards. Tomorrow, we will find them also in business cards, note pads, desks, keys, doors, shoes, and wall paint. The result is an intelligent environment that integrates computing seamlessly into work and life.

The components in such domain-specific architectures are generic and interact with one another through standardized interfaces termed interconnections. As needed, the generic components may be parameterized to simplify customization for particular applications. Specific implementation components can be saved in and retrieved from a repository, and advanced technology may be used to generate particular implementation components more or less automatically from requirements. This approach to composing application systems from a collection of modular components and standardized interconnections is called domain-specific software architecture , ^[2] and the most frequent reusable multi-part structures are termed patterns . ^[3]

^[2] http://www.htc. honeywell .com/projects/dssa/dssa_whatis.html

^[3] http://www.enteract.com/bradapp/docs/patterns- nutshell .html

Behind this simple idea of reusable architectures is a range of assumptions and beliefs that provide a fuller context and rationale. Elements in this context include the assumed development process, the key resources that are required, and the principal sources of leverage. Furthermore, it is necessary to understand the key constraints, sources of variation and controls, and flexibilities and exceptions. You also need to understand relationships to other methods and standards. Each of these is discussed briefly .

Computing devices already cover a wide range of platforms, computing power, storage capacity, form factors, and user interfaces. We expect this heterogeneity to increase over time rather than decrease, as new classes of devices such as pads or car computers become widely used. Today, applications are typically developed for specific classes of devices or system platforms, leading to separate versions of the same application for handhelds, PCs, or cluster-based servers.

Furthermore, applications typically need to be distributed and installed separately for each class of devices and each processor family. As heterogeneity increases , developing applications that run across all platforms will become exceedingly difficult. As the number of devices grows, explicitly distributing and installing applications for each class of devices and each processor family will become unmanageable, especially in the face of migration across a wide area. Therefore, a new computing architecture needs to be put in place.