The Stupid Network versus Intelligent Network debate rages on. Ostensibly, the debate is about creating a new kind of network, a network unencumbered by outdated assumptions; a network capable of starting afresh, free to explore new paradigms. And yet the tendency to define the new (or Stupid) network as something in opposition to the old (or Intelligent) network admits much that is itself dated.

If the frustrations occasioned by the existing network are old, so too are the expedients meant to compensate for the existing network’s limitations. So, if the Stupid Network were no more than the ultimate compensation for the old network, how new would it be, really?

COUNTERPARTS, IF NOT COMPLEMENTS

The partisans in the Stupid versus Intelligent debate tend to fall into two camps. In the first camp, we have people with the user’s perspective; in the second, people with the provider’s perspective, experts practiced in designing, building, maintaining, and regulating networks.

Once, the most compelling example of networking was the public switched telephone network (PSTN), and experience with the PSTN indicated that designing, building, and running a network required a fair amount of ingenuity and large amounts of capital. Data communications were provided through either open or proprietary solutions. Carriers and service providers made substantial investments to deliver successful and ubiquitous services. They had to design and build the switching and transmission infrastructure, as well as provide the operations, management, administration, and provisioning services needed to keep it running.

It was difficult, time consuming, and expensive to introduce new services. Furthermore, the network provided only basic transport and telephony services. Anything beyond this was up to the users to provide between themselves.

Not surprisingly, providers and users both found this state of affairs frustrating. The network evolved, however, because processors and memory became more powerful, and digital technologies advanced over analog technologies. New bandwidth became available; new protocols were defined. Services running on switches moved to platforms separate from the switches. This was the beginning of the Intelligent Network.

At the same time, people using the network began bundling technology to create new services at the desktop based on the assumption that the network provided basic transport and telephony services. This was the beginning of computer telephony.

To date, progress has been slow on both the Intelligent Network and the computer telephony fronts. A thorough examination of why this should be so is beyond the scope of this article; however, it may help to keep the user’s perspective in mind, to consider, from the user’s point of view, what makes applications and services attractive and successful. Basically, users want applications and services that are easy to use, widely and readily available, inexpensive, and reliable. If neither the Intelligent Network nor computer telephony has produced a "killer app," it may be that neither has produced an application or service that has met all these criteria.

ENTER IP: CONVENIENCE OR CHALLENGE (OR BOTH)?

Whenever a new technology comes along, it can stimulate at least two kinds of responses: It can be seen as an elaboration on an existing pattern (which can be convenient), or it can be seen as something with the potential for elaborating patterns of its own (which can be a challenge). Occasionally, both responses are justified, as is the case with the Internet.

The Internet, unlike the PSTN, is based on connectionless, packet-switched protocols. These protocols are known collectively as the Internet Protocol (IP), which is independent of the underlying physical media and the protocols associated with these media. IP, thanks to these properties, essentially expands the traditional scope of the network. Users now have the option of controlling the network. Infrastructure can grow incrementally and doesn’t require the same, initial massive investment as the telephony network.

IP’s potential for expanding the scope of the network was described in memorable terms by Internet essayist David S. Isenberg, who introduced the idea of the Stupid Network in his essay, "Rise Of The Stupid Network: Why The Intelligent Network Was Once A Good Idea, But Isn’t Anymore. One Telephone Company Nerd’s Odd Perspective On The Changing Value Proposition" (www.isen.com). The Stupid Network, in theory, makes the network easier to use with features and optimizations extrinsic to the network. The Stupid Network may allocate more bandwidth, memory, and processing power to give users the ability to monitor and control the network. There is also no controlling authority such as a carrier or service provider, and the Stupid Network is broadly specified.

The Stupid Network, in its emphasis on the user’s role, and its tendency to diminish the service provider’s role, superficially resembles computer telephony. Or, it at least seems to be in sympathy with the "do it yourself" spirit often expressed by traditional, desktop-oriented computer telephony. But is the Stupid Network, or any other IP-inspired scheme, no more than an opportunity to replay an old debate, albeit on a grander scale? The answer may depend on your point of view.

DIFFERENT OUTLOOKS, DIFFERENT GRAPHS

All network models — including the Intelligent Network, computer telephony, and the Stupid Network — are really different graphs using the same basic network elements, such as physical and virtual nodes and connectors. Physical nodes are real devices and equipment; virtual nodes consist of more abstract elements such as programs and their information. Connectors are the channels, such as fiber, radio, or even Inter- process Communications (IPC) mechanisms of an operating system, that enable communications between the nodes.

In any of these networks, the development and deployment of a service depends on the relationships between the nodes and connectors. Two graphs can represent these relationships. One graph represents where these components reside physically (that is, customer premises, access exchange, carrier, service provider, etc.); the other graph represents where the components reside virtually (that is, what are the interactions between components and who controls and monitors them). The physical and virtual graphs for the same service can be very different from each other, although each is correct.

The people that design, build, use, run, and regulate the network have different requirements that drive the organization of these graphs. During the early history of the network, the people who designed, built, and ran the network were responsible for putting these pieces together. Although this still happens, more users are becoming involved in these activities. Not all users, however, will be able, or want, to have this sort of involvement either because of limited experience and/or knowledge.

In the future, as feedback mechanisms such as collaborative filtering are introduced and the semantic content of information becomes available through languages such as the Extensible Markup Language (XML), the network itself will be able to organize information and relationships between nodes.

PERPETUAL DECONSTRUCTION AND CONSTRUCTION

The common thread during the history of the telephone network and the Internet is that as simple elements and concepts cross thresholds of complexity, they organize and become new abstractions. These new abstractions absorb the complexity of the previous elements and become the foundation for future services. At some point, these abstractions typically become insufficient and quite often deconstruct and then reconstruct with new elements and concepts that have since appeared to become, yet, new abstractions.

As an example, the future integration of Internet telephony with telephony services will require both the further deconstruction of Intelligent Network and computer telephony platforms and reconstruction with Internet elements such as servers, gateways, and gatekeepers, as these various platforms move toward a unified service architecture.

CONCLUSION

Someday, the telephone network and the Internet will become one network, called the Network or the Net. The definition of this Network — whether stupid, intelligent, or something else — will depend on the perspective of the person talking about it.

Intelligence — that is, the programs and information — will reside in many different places across the network, including the customer premises either at the desktop or within a data center, at an edge switch, or within service platforms. If the intelligence physically resides with the user, it is the so-called Stupid Network. If it resides with the carrier or service provider, it is the so-called Intelligent Network. In either case, the network will use similar network resources.

The stupid and intelligent labels do not do justice to the complexity and diversity of the network. These labels also can create mental boundaries to future possibilities. Some of the confusion arises from the use of the word network. "Network" is typically used to describe the physical nodes and connectors monitored and controlled by the carriers or service providers. If we accept this usage, the Stupid Network is a concept that makes sense.

If the usage of "network" is broadened to encompass everything between users of a service, including devices, equipment, programs, or information, then the Stupid Network is only part of the picture. There are certain applications and services that need the ability to find people or information on a global scale. This information in many cases will only be maintained and available from the carriers or service providers and be part of what has been called the Intelligent Network.