The Intelligent Wireless Web: Our Next Generation Web
by H. Peter Alesso, www.web-iq.com
If we concentrate really hard, would it be that difficult to envision the future of the Web?
Perhaps it would just take some basic knowledge, a little imagination, a dash of adventure, and pinch of insight. We might find that it is as easy as one, two, three:
1. We merge the Next Generation Internet (NGI) with Internet2,
2. We experiment with interactive intelligent programs, and
3. We improve the user interface with speech recognition, while extending connectivity through wireless devices.
That's the future of the Web - a combination of broadband delivery, innocuous interfaces, and ubiquitous access - and all with interactive intelligence. Eventually, all this is not only possible, but highly likely.
However, foreseeing such an end point 3 to 7 years in the future, is one thing, and developing a credible scenario for achieving that result is something else. Let's try to put together some of the pieces to fashion the necessary credible scenario.
1. Merging Next-Generation Internet (NGI) and Internet2
The Next-Generation Internet (NGI) initiative is a multi-agency Federal research and development program that is developing advanced networking technologies, developing revolutionary applications that require advanced networking, and demonstrating these capabilities on test-beds that are 100 to 1,000 times faster, end-to-end, than today's Internet.
The key distinction between the NGI initiative and Internet2, is that NGI is led by and focuses on the needs of the federal mission agencies, such as DoD, DoE, NASA, NIH and others, while Internet2 is university based through grants.
However, the NGI program focuses on some of the same emphases as Internet2:
Advanced infrastructure development (i.e., networks that can perform at much greater levels than today's commercial Internet).
Advanced applications development.
Research into technologies that will enable advances in infrastructure and applications.
Currently, the Federal NGI initiative and the university-led Internet2 are working together. The National Science Foundation (NSF) has made more than 70 High Performance Connection awards to Internet2 universities. These merit-based awards allow universities to connect to NSF's very high performance Backbone Network Service (vBNS). vBNS connectivity is a key part of NSF's NGI program.
Internet2 universities are establishing gigaPoPs (Gigabit per second Points of Presence) that provide regional connectivity among universities and other organizations. Through the gigaPoPs, universities will connect to NGI networks and other advanced Federal networks, including the vBNS, NASA's Research and Education Network (NREN), DoD's Defense Research and Education Network (DREN), and the Department of Energy's Energy Sciences network (ESnet). The NGI and Internet2 will help ensure that advanced networking services are available on interoperable backbone, regional, and local networks that are competitively provided by multiple vendors.
Progress in this area is proceeding rapidly and we are likely to see significant results in the next 3 to 5 years.
2.0 Experiment with interactive intelligent programs
For the most part, the Web can be considered to be a massive information system with interconnected databases and remote applications providing various services. While these services are becoming more and more user oriented, the concept of smart applications on the Web is still in its infancy. So, how will adding intelligent agents, smart applications, and Artificial Intelligence (AI) programs to Web sites, contribute to the development of the Intelligent Web?
To begin to address this issue, we will have to explore some uncharted territory and face some probing and provocative questions, such as:
How smart are today's Web applications?
Information Portals are one of the most sophisticated applications on the Web today. During 1998, the first wave of Internet Portals became very popular. They provided consumers with personalized points of entry to a wide variety of information on the Internet. Examples included; MyYahoo (Yahoo), NetCenter (Netscape), MSN (Microsoft) and AOL.
Subsequently, Enterprise Information Portals (EIP), also called Corporate Portals, provided ready access to information over Intranets and the Internet. Corporate Portals moved beyond the delivery of information, they also provide a way to integrate the many disparate systems and processes that are typically used within an enterprise. Corporate Portals are able to use XML to integrate previously separate legacy systems and to provide a single point of entry to these processes1.
In addition, EIPs now act through state-of-the-art (AI) applications to search, retrieve and repackage data for access centers that tie together people and data. They link e-mail, groupware, workflow, collaboration, and other mission-critical applications. The EIP is in the process of developing into an even more powerful center through component-based applications called Web Services. Web Services use XML standards, frameworks and schema that make up today's most sophisticated applications.
But considering the incredible amount of programming, installing, debugging, and maintenance they require, would you categorize any of these inflexible programs, as truly intelligent?
What is Web intelligence?
Intelligence usually refers to the ability to reason, to solve problems, to remember information, or to learn new ideas. In May 1997, IBM's Deep Blue Supercomputer played a defining match with the reigning World Chess Champion, Garry Kasparov. This was the first time a computer had won a complete match against the world's best human chess player. For almost 50 years, researchers in the field of artificial intelligence had pursued just this milestone. The success of Deep Blue and chess programming was important because it successfully employed both types of AI methods: logic and introspection.
For now, the Web consists primarily as a huge number of data nodes (containing texts, pictures, sounds, and video). The data nodes are connected through hyperlinks to form `hyper-networks' that collectively could represent complex ideas and concepts above the level of the individual data. However, the Web does not currently perform many sophisticated tasks with this data. So far, the Web does not have some of the vital ingredients it needs, such as, a global database scheme, or a global error-correcting feedback mechanism, or a logic layer protocol, or a method of adopting Learning Algorithms. As a result, we may say that the Web continues to grow and evolve, but it does not adapt. And adapting is an essential ingredient of learning.
Well, the jury may still be out on defining the Web as intelligent, just yet, (and may be for some time), but we can still consider ways to change the Web to give it the capabilities to adapt and therefore to learn.
How will the Web get smarter?
Central to human intelligence is the process of learning or adapting. Likewise, machine learning may be the most important aspect of Artificial Intelligence (AI), including behavior, cognition, symbolic manipulation, and achieving goals. This suggests that AI software should be concerned with being changeable or adaptable. The challenge for AI is to learn capabilities for helping people derive specifically targeted knowledge from diverse information sources, such as, the Web. Subsequently, one of the challenges facing Web Services includes developing a global consensus for an architecture that lets applications (using object-oriented specialized software components) plug into an "application bus" and call a Web AI service.
We can defined a Learning Algorithm as a process that takes a data set from a database as input and after performing its algorithmic operation returns an output statement representing learning. As Web increases the percentage of applications and protocols with Learning Algorithms, we can expect improvements in performance both in quality and type.
The Web may become a learning network through components of AI agents and AI application built with adaptive software languages and connected by Web AI Service Portals.
However, regardless of how AI applications are processed on the Web, a vital challenge will be the establishment of trusted information. The process must build trust of information and will include a form of Information Registration and Validation. This will remain an issue for some time to come.
Mobile commerce products are now being integrated with wireless Enterprise portals. The m-Commerce products are designed to work efficiently within customers existing suite of mobile products and services. The m-Commerce products range from Auto registration capabilities, Universal cart & catalog, to e-Wallet product.
The road map for achieving a set of connected applications for data on the Web in the form of a logical web of data is called the Semantic Web. An underlying idea of semantic networks is the ability to resolve the semantics of a particular node by following an arc until a node is found with which the agent is familiar. The Semantic Web, in competition with AI Web Services, forms a basic element of the Intelligent Web (see Figure 1).
Figure 1. Web Architecture - From dumb and static to intelligent and dynamic
Finally, whether learning is achievable from AI Portals at all, remains extremely controversial. The virtue of controversies is however, that they
motivate experts into uncovering dormant capabilities in response to challenges to resolve competing paradigms.
For example, opportunities for wireless developers and Internet service providers will greatly expanded when they are able to reach all mobile users by developing infra structure that is able to support; any wireless carrier, any wireless network (TDMA, CDMA, etc.), any wireless device (pager, digital cell phone, PDA), any wireless applications, any Web format (WML,
etc.), any wireless technology (WAP, SMS, pager, etc.), and any medium (text, audio, text-to-speech, voice recognition or video).
Trying to accomplish universal interoperability is a demanding challenge requiring resolution of the competing paradigms; balancing diverse proprietary standards (which
stimulate competition) against open standards (which offer universal access).
Progress in this area is proceeding and we are likely to see significant results in the next 5 to 7 years.
3.0 Improving the User Interface and Extending Connectivity with Wireless Devices
Imagine living your entire life within the confines of a specified region that surrounds you. You could call this region, your Personal Space. As you travel from home to work this designed region travels with you just like a 'bubble.' If you look around this space, how many electronic devices would you see? How many wires would exist? With every new electronic device, you add to the 'cable tangle' around you both at the office and at home. But, now, wireless technology can add connectivity to these devices without the encumbering tangle.
Wirelessly connected devices create a network infrastructure called a Wireless Personal Area Network (WPAN). The obvious application of a WPAN is in the office workspace. With this technology, your essential workspace electronic devices will be wirelessly networked together. These could include your desktop, mobile computer, printer, handheld device, mobile phone, pager, etc. Your personal devices could, for example, wirelessly update your appointment calendar on your office PC. You would have greater flexibility in arranging your office because peripherals would no longer need to be within cable length of the PC. The growth of home automation and smart appliances could also use WPAN applications, just as in the office.
WPAN will also allow devices to work together and share each other's information and services. For example, a Web page can be called up on a small screen, and can then be wirelessly sent to a printer for full size printing. WPAN can even be created in a vehicle via devices such as wireless headsets, microphones and speakers for communications. Additional, wireless devices may eventually be embedded throughout public places to provide continuous connectivity as you travel within your Personal Space from one location to another.
In today's environment, information (such as that available from AI Web Service Providers) is one of our most valuable commodities and small, cheap and yet powerful devices, may offer universal accessible to vital information. Thus a lifetime of knowledge may be accessed through gateways worn on the body or placed within our Personal Space.
As envisioned, WPAN will allow the user to customize his or her communications capabilities enabling everyday devices to become smart, tether-less devices that spontaneously communicate whenever they are in close proximity (see Figure
Figure 2. Personal Connectivity - From Wired to Wireless
With billions of various devices already in use today, developing multipurpose communications that can receive and transmit compatible signals is a daunting challenge. At the local level, Personal Area Networks (PAN) form device-to-device interfaces at work and at home. At the global level, we must adapt an interlacing complex of networks (such as integrating NGI and Internet2) to connect compatibly to a large number of possible device-to-device combinations.
The key problems with the small devices available today are; their screens are small with low resolution, their power and memory is limited, and their bandwidth inadequate.
Small mobile wireless device computing environments are not able to run large, complex operating systems and applications. Instead, distributed applications, which gain their capabilities from collections of separate devices working in concert, will be necessary. Unlike desktop computers small mobile wireless devices use a variety of processors and operating systems and are programmed in a variety of languages.
One solution to output problems may be larger screens. The extra space could come from a flexible screen that unfolds like a map. Plug it into a pocket PC and you have a workable product. But pocket sized foldable screen technology is still a few years away. An alternative is "electronic ink" technology being developed at E. Ink of Cambridge MA. Electrostatic charges orient white microscopic particles suspended in tiny spheres. Unfortunately, electronic ink is also several years from practical use. Another approach keeps the display small but offers good resolution using magnifying lens mounted on monocular units or goggles. Sony's Glasstron and Eye-Trek from Olympus both give the viewer an image equivalent to a 132-centimeter screen seen from two meters away.
If output over small screens looks troublesome, input problems are even harder. Just think what it's like using keypads from your cell phone to send typed messages. Certainly several cellular-phone manufacturers, including Motorola and Nokia, are trying fledgling speech recognition already in the form of simple "yes" or "no" responses, or in the form of one-word names of stored phone numbers.
Speech recognition and speech synthesis offer attractive solutions to overcome the input and output limitations of small mobile devices, if they can overcome their own limitation of memory and processing power through the right balance for the client-server relationship between the small device and nearby embedded resources. The essential components for achieving this balance are new chip designs coupled with open adaptive software. The new chips may provide hardware for small devices that are small, light weight, and consume little power while having the ability to perform applications by downloading adaptive software as needed.
The success of mobile communications lies in its ability to provide instant connectivity anytime, anywhere in a practical and user-friendly manner. If the convergence of the mobile wireless and fixed information networks is to have significance, the quality and speeds available in the mobile environment must begin to match those of the fixed networks. How to build this broadband wireless network is the difficult question. Telecom companies will need to spend billions of dollars to catapult today's narrowband (9.6 kbps) cell-phone infrastructure to achieve broadband capabilities.
Working against broadband access is a fundamental law of data communications. Back in 1948, Claude E. Shannon of Bell Labs, found that the maximum amount of data that can be transmitted through any channel is limited by the available bandwidth (the amount of radio-frequency spectrum it occupies) and its signal-to-noise ratio (the signal to be communicated versus the background interference). The need for high-speed data services and high quality voice transmission under roaming conditions represents significant challenges for wireless communications.
The invasion of digital communications into the wireless world is already in progress. Analog cell phones were found to be useful as a tool, but it is digital phones that have become a mainstay of wireless communications throughout the world.
Today, you can buy a book from Amazon.com, reserve tickets for a concert, or access your company's intranet right from your mobile phone. But technical limitations make it a tedious task. Wallets, such as, Microsoft's Passport and Yahoo!Wallet, simplify and speed up data entry by automatically sending the pertinent information to an e-tailer when a transaction is complete. However, mobile-commerce is more attractive when viewed from the perspective of a longer time horizon.
There are several mobile competitors influencing different regions of the world. The most widely used cellular network technology is GSM (Global System for Mobile Communications), a Time Division Multiple Access (TDMA) system in both Europe and Asia. Unfortunately, TDMA is less adaptable to the Internet's bursty data flows. A key alternative, Code Division Multiple Access (CDMA), faces strong opposition in many quarters.
The cellular phone industry has been unsuccessfully trying to market wireless packet data services to consumers since 1994. At that time, the industry offered packet data overlay on top of the AMPS analog cellular service. But due to bandwidth limitations (i.e. the maximum data throughput with CDPD is approximately 14.4 kbps), wireless data failed to catch on with US consumers.
More recent technologies, such as HDML, WAP, Compact-HTML and J2ME have helped solved part of the bandwidth issue by reducing the amount of information sent over wireless, and new wireless packet data technologies (e.g. GPRS, IEEE 802.11) are increasing bandwidth. There remains, however, the problem of what to do when the mobile subscriber unit roams to a different IP subnetwork. Mobile IP offers a solution to this problem by having the router on the home subnetwork "tunnel" IP packets to the mobile at a "care-of address" on the new IP subnet-work. There are also some technical issues to be worked out with Mobile IP, such as firewalls.
Mobile IP represents a simple and scalable global mobility solution, but lacks support for fast handoff control, authentication, real-time location tracking, and distributed policy management that is found in cellular networks today. In contrast, third generation cellular systems offer seamless mobility support, but are built on complex connection-oriented networking infrastructure that lacks the flexibility, robustness and scalability that is found in IP networks. Future wireless networks should be capable of combining the strengths of both approaches without inheriting their weaknesses.
Progress in this area is proceeding and we are likely to see significant results
during the next 3 to 7 years.
Our vision for the next generation Web is a result of combining some basic knowledge, a little imagination, a dash of adventure, and pinch of insight to expose three basic steps:
1. We proposed merging the Next Generation Internet (NGI) with Internet2,
2. We proposed experimenting with interactive intelligent programs, and
3. We proposed improving the user interface with speech recognition, while extending connectivity through wireless devices.
That's our future of the Web - a combination of broadband delivery, innocuous interfaces, and ubiquitous access - and all with interactive intelligence within the next 5 to 7 years.
Whether AI (artificial
Intelligence) Web Services will eventually be characterized as contributing to Web "thinking," is considered to be a relatively unimportant question, because whether or not "smart" programs "think", they are already demonstrating that they are useful. The discussion of how to build intelligence to enlighten the optical pathways that inhabit the Web is foreshadowed by the evolution of mobile wireless communications through the efforts of today's innovators.
Alesso, H. Peter, and Smith, Craig F., The Intelligent Wireless Web, Addison -Wesley Professional, ISBN: 0201730634, Dec. 2001.