HOUSEHOLD ARCHITECTURE ISSUES

The architecture of a home network presents unique challenges that even complex business networks have yet to address. Three primary forces will govern the household superstructure architecture:

• Structure variance— The configuration of ENDs will vary greatly from home to home.

• Usage variance— The application use for each END will vary greatly.

• Client-server dynamic— A network architecture will be comprised of multiple client-server relationships across all levels of network environment. The household architecture issues are important because they relate to the technical requirements that the Tier 1 infostructure must support.

Structure Variance

While the Tier 1 infostructure (Internet backbone and home infostructure) provides a universal technical architecture for all consumers, the Tier 2 ENDs superstructure will vary greatly with respect to the needs and preferences of each household. This is why industry-wide standards for the Tier 1 infostructure will be so important. It is essential to establish robust yet flexible technical specifications for substructure technologies. These specifications must optimize the substructure to ensure interoperability and consistent functionality for the multitude of ENDs that will occupy the consumer space in an infinite variety of combinations and configurations.

Because no two households will own the exact make, model, and quantity of ENDs, the superstructure will vary greatly across the demographics of the home user market. The technical features of the ENDs and the configuration architectures within each home will prove much more unpredictable than business network environments. Developers of the supporting substructure systems (Tier 1) must be aware of this fact. This represents one of the most important conceptual requirements for the home network, especially since it must operate in an automatically configured "plug-and-play" way.

Usage Variance

Not only will the physical superstructure vary greatly between homes, but the applications and usage patterns within the home will prove equally unpredictable. This is relevant when comparing the computing usage of traditional business environments with the possible usage scenarios for a networked home.

Business environments consist of straightforward network applications that workers use in a fairly consistent and homogeneous nature. For example, common network applications for business users include Internet browsing, database queries, e-mail, multimedia correspondence, and file sharing. In contrast, the usage patterns in the next generation of home networks will be much more complex and unpredictable. Consider that any END sold to a consumer might have embedded intelligence and network communications capabilities. For example, a manufacturer might sell a networked washing machine with embedded intelligence that will sense malfunctions, automate repair servicing, and synchronize its operations schedule with the electricity provider for optimal pricing.

In most cases, the consumer will be unaware of the technological underpinnings of such a system. The purchaser will not consider that "doing their laundry" also entails running "computer applications." They only care that the washing machine offers valuable new features that will lower utility bills, increase the ease-of-use, and assist during malfunctions by automatically dispatching a service request.

This scenario conveys the fundamental value that the pervasive computing vision offers, where technology and computing give way to the tangible benefits and practical forms of consumer value (like saving money and simplifying work).

Imagine if this scenario is extended to the entire spectrum of ENDs, including all of the appliances, electronics, and computers within the home. Also consider that the variety and complexity of these user applications will increase over time as OEMs leverage them to differentiate their products and achieve continual competitive advantages. Although the possible outcomes of this trend are not within the scope of this discussion, substructure developers must consider the implications of usage variance during their development phases.

Client-Server Dynamic

In their evolution toward a distributed computing environment, the design and function of ENDs will eventually reflect a client-server dynamic. While this architecture may resemble the client-server system of business environments, it will be redesigned to accommodate the unique needs of pervasive home computing. New devices will be sensitive to cost and to the limited technical know-how of the home user. The distribution of digital content from source devices (front-ENDs) to the rendering client (back-END) is a distinguishing feature of evolved devices. In a sense, back-END devices are the servers, while front-END devices are the clients.

The client-server dynamic achieves greater computing efficiency by centralizing shared resources into single, multipurpose devices. These devices process and distribute digital content to a multitude of thin clients, creating a cost-effective technology solution that scales easily depending on individual consumer needs.

The client-server dynamic will engender multiple client-server relationships at different levels across the network environment. In some cases, the client device might be a platform-independent END that renders content. But other devices might play the dual role of both client and server, depending on their technical relationship to other devices.

For example, a home music server, which might become a future standard for storing music titles (replacing your CD collection), could act as both a client and server in certain network architectures. Consider a scenario where the music server interfaces with the residential gateway to download titles from an Internet-based music vendor. Then the server stores the files directly to a storage device. The music server could, in turn, support multiple speakers by connecting to a powerline-based IAN. The powerline IAN would supply both the power to the run the amplifiers (embedded in the speakers) and the data for the music titles.

Music would be streamed on demand (through a touchscreen control interface) from the server, via the IAN, to multiple networked speakers.

The streamed data would travel between the devices via the AC electrical connections using powerline Ethernet transceivers. In this way, the front-END speakers could render the sound data through an embedded decompression engine that corresponds with the data storage format of the back-END music server (most likely MP3 or other compression technology). Figure 1.3 shows this conceptual architecture where the speakers and touchscreen take the form of next-generation front-ENDs. The home music server takes the form of a back-END, which interfaces with an Internet-based commerce server via the residential gateway, to download newly acquired titles. This model demonstrates the possibility for new architectures that developers could achieve when considering the home infostructure during the design processes of next-generation ENDs.

An alternative to this architecture is for music content to stream from an Internet-based source directly to the home's front-ENDs (thin-client speakers) via the home infostructure. In reality, both of these architectures may be practical entertainment applications for the home.

The client-server dynamic simply dictates the fundamental principle of efficiency for network computing environments. The distributed media system poses the possibility for a more efficient architecture by inserting the home music server between the Internet servers and the front-END speakers.

This creates a redundant, multilevel client-server architecture. In this case, the redesign involves decoupling the distribution media (CD or MP3 source) from the display, control, and playback technology, using the IAN to interconnect them in a distributed computing architecture.

In this case, a distributed computing architecture is more efficient because it reduces unnecessary traffic that may cause a bottleneck on the Internet, or through the home infostructure, each time a user wants to hear a song. It may be better to use the home infostructure for single-use downloads to a local device, like the back-END home music server, which then supports the user's media-rich application needs, using the higher-bandwidth features of the IAN.

Ultimately, the redundant client-server dynamic cannot determine which architectures will prove most viable in the marketplace. However, the concept can be applied as an analytic tool to help determine how to achieve greater efficiencies at multiple levels of the network architecture.