Visions: Application and Technology Drivers

Introduction

This section of the Internet Tool Survey sketches a number of visions of the future that will be enabled by Internet-Web-Object-Middleware-DBMS integration. The first several visions are application visions, that is, visions of what this integration will mean to user communities. The rest are technical visions of how enabling technologies may change. Internet, web, object, middleware, and DBMS integration are not the only enabling technologies in many of these visions but they are important ones in providing the "glueware" for connecting applications. At the level of this paper, details of how specific technologies enable these visions are not included; many of these enabling technologies are covered in other sections of the Internet Tools Survey.

Application Drivers

Information Super Highway - Networked Community

This is the overarching vision. Its thesis is that our society will fundamentally change as access to information resource tools becomes pervasive. The scale is global with high quality information accessible (almost) everywhere and anytime. The thesis is that technologies like the Internet and WWW provide major enabling connectivity, and open the door to new networking technologies (ATMs, cable, ...); entertainment, publishing, and computing will merge (as TVs, phones, and Internet merge); structures of organizations are likely to change (virtual offices, virtual enterprises); information sources will become active (agents) and simulations more interactive, realistic, and collaborative (virtual reality); and all this will affect our economy (electronic commerce, electronic mall), our politics (all major candidates now have homepages), our defense (information warriors, digital battlefield), our education (digital libraries), and our lives (virtually everything will be connected).

Electronic Commerce

Secure electronic cash, micropayments, watermarked copyrights, EDI - the Internet will be the gateway to selling direct electronically to the home or office. This is probably the vision that will fuel the most innovation - a modern gold rush. Electronic commerce will change the way we buy goods and services. Currently in the U.S. (but not in Europe), consumer credit laws hold individual credit card holders harmlessagainst fraud, making it OK to send credit cards over the Internet-but merchants must be cautious about accepting those credit card numbers because of that fraud. Electronic cash is just the start. The DARPA-backed CommerceNet Consortium is one of the prime movers in this area.

The Internet is an open marketplace where buyers and sellers meet. On-line shopping and catalog sales provide a means for merchants to monitor browsing and buying habits of individuals and groups, and provide specialized services (Virtual Vineyards, flowers, automobiles, buying services, virtual malls) based on profiles and preferences, giving rise to direct, targeted sales and narrow-casting. Service industries such as real estate, insurance, and travel are showing up on the net. Newspapers, magazine, and books are sold over the net. Many store fronts only exist on the net. Virtual books sold over the net for $1 might provide greater royalty than physical books sold in stores for $10, and individuals can themselves control production. Current information can be up-to-date and customized for consumers. So, instead of buying a generic car (or tank) manual, you buy access to an electronic manual that is customized by a car's specifications, updated as work is done to the car (or as recall notices are issued), reminds the owner when to schedule preventive maintenance, contains information on local service providers, and provides just-in-time information on home-auto repairs (or home printer repairs).

New markets for information products that formerly cost too much to package are possible with micropayments: licensing fragments from articles, the use of a program by the minute, locator and information summary services, a few minutes of expertise.

Advertising on the net provides not only ways to reduce costs for information providers but also gateways to in-depth sales information or direct connections to a sales force from interested parties. Consumers will eventually be able to compare value of products more easily or order custom products delivered to their home.

Incentives are high and there is a manifest destiny that information services will become increasingly sophisticated. Paralleling this will be the need for increased infrastructure standards as pervasiveness increases. A CommerceNet/Nielsen study shows that 11% of the US and Canadian population ages 16 and older has used the Internet in the last 3 months.

Intranets and Enterprises

Yesterday, the "Web" and the "Enterprise" were two vastly different worlds; tomorrow, they will coalesce; today, both are in quick transition. Several changes are occurring in corporate computing:

Intranets -- For the past several years, workgroups have been adopting workstations, LANs and a host of Internet tools, gradually displacing expensive, centrally controlled in-house or proprietary corporate networks with Intranets, that is, Internet-based networking solutions used within an organization but separated from the outside by firewalls. In May 1996, Hewlett Packard, the second largest computer maker in the U.S., turned off its last mainframe (savings projected $8M in yearly maintenance costs) and completed conversion to a common PC Intranet environment (savings projected $25-30M annually) and is preparing employees for "virtual offices and knowledge networks." Intranets are easier to maintain, less costly, less proprietary, and more decentralized than proprietary solutions. Corporations and communities such as DoD's Intelink community use this kind of solution for secure internal communication (Intelink is an Internet-based system created to enable DoD intelligence agencies to share information). Financial institutions, the travel industry, real-estate, and several other communities use specialized private networks for secure or controlled communications. While pervasive today, these physical divisions of networks and Internets into public and private physical networks does not serve all needs: large organizations are at risk from a breach of security via dial in access; some communities, such as medical networks, are inherently distributed and yet need secure communication over insecure channels; and, increasingly, financial transactions are occurring over the Internet. Tremendous effort is currently being put into secure communications and solutions that support reasonable risk models are emerging. Near-term, approaches to cross corporate firewalls when needed with push-pull data and computation are coming on line. Virtual private networks and multiple virtual enterprises on a desktop are moving specialized secure networks with discretionary access control onto the insecure Internet.
Better information flow -- email, newsgroups and the Web have made communication across corporate boundaries as easy as within corporations. This leads to a flow of ideas and a greater cooperation as corporations in infrastructure areas (e.g., Internet protocols, middleware, DBMS, workflow) realize they need to cooperate to insure interoperability. The Web enables "thin clients" at remote locations (globally) to access and view data formerly limited to small workgroups.
Customer-centric solutions -- Companies will open up and integrate themselves with their customers and suppliers over the Internet. A much closer relationship with customers is possible so companies are making back office information (e.g., order entry, design, schedules) progressively more visible both within and outside corporate firewalls in an attempt to make information more useful to more people. Back-office applications will be reengineered to allow internal corporate applications (e.g., so employees can review their personnel files and benefits from their home and employees on sick leave or vacation can stay connected as needed) as well as outside interaction (e.g., so customers can track their orders being processed).
Componentware Infrastructure - today's two-tier client-server architectures often involve a generic server and a fat client since application logic is typically stored on the client. This is effective in isolation but increasingly the functionality needed is located in multiple servers on the network. Web-based solutions provide thin clients with application logic downloaded as needed or business rules common to many applications stored in a different server than the generic server (a three tier scheme). Functionality (e.g., directory services, security services, replication, versioning, business rule reuse) now found bundled differently within different software packages is becoming componentized and reusable with different DBMSs, KBMSs, CASE, workflow, and other tools. Closed and monolithic tools will be increasingly less-competitive since no single vendor will be able to supply best-of-class single-vendor turnkey solutions for all components.
Dynamic and Continuous Change - as a business (or military scenario) changes, the software to support it must change to reflect the changes. When middleware and business rules are separated from application logic, they are easier to maintain or change as the organization's needs evolve. Internet and Web-based solutions can reflect changes immediately throughout the organization when central MIS is aligned with business changes.
Scaleable solutions - Forrester Research has reported that today two-thirds of all client-server applications are used within a single department. Web-based solutions mean these same applications can more easily scale to the enterprise.
Software deployment and remote maintenance - Instead of installing all application binaries on each computer to be served, application software can reside on the network, be centrally maintained, and downloaded as needed. An organization can deploy software to one user, groups of users, an enterprise, or a community without having to explicitly fanout, install, and test the software on every client computer. Trouble-shooting will be from remote locations. This kind of scaleability can increase return on investment on software development by orders of magnitude, and also means lower cost per user. Gartner Group studies indicate that in a five year period, 75% of the real cost of client/server applications is spent on deployment and maintenance.
Organizational changes - the Web levels organizations since they all can advertise services equally allowing small organizations to compete with large ones. Organized information spaces can reduce the need for layers of reporting and can marshal the knowledge of an organization and make information more accessible, more valuable, more available to all, and can make an organization more responsive.

Virtual Enterprise and Virtual Office

The structure of the traditional office or enterprise will change if information and communication makes remote virtual presence the same as "being there." Virtual offices and virtual enterprises are legal entities that make heavy use of distributed infrastructure technologies to support remote work.

A virtual office (VO) is a distributed office where all or almost all office functions are distributed. It is a step beyond telecommuting, and relies on good organization and good electronic connectivity for communication. Office personnel may or may not be very mobile. Commuting is reduced and hiring can be best-of-class without geographic barriers. Technology enablers are email, WWW, video conferencing, phone lines, sharable files and compound documents; increasingly, much of the same groupware offered in the central office will be available to collaborators working anywhere in the world. OBJS is an example of a virtual office.

A virtual enterprise (VE) is an enterprise where the members are organizations that contribute certain resources to further the objective of the organization. Some VEs are ephemeral, created for a certain objective; others may have long life. Good interoperable computing infrastructure is needed to enable both virtual offices and VEs. The NIIIP Consortium is an example working to integrate Internet, Web, OMG, workflow, KBMS, agent, and STEP technology for this application.

Command and Control - Information Logistics - I3

In business or in a quickly unfolding international crisis, it is necessary to collect "interesting" information about a situation quickly. How can a swarm of agent-objects be released to locate and filter un/semi/structured information quickly?

Intelligent Integration of Information (I3): Given the huge reservoirs of information locked in legacy systems, how can that information be located, extracted, and merged with other information? Wrappers, mediators, glue and agents provide a family of approaches.

Situation assessment: A commander at a remote location views a map of a specific area and gets summary information on the events of interest occurring in that area, which may be stored or analyzed at other locations. In general, a situation description is a typed, structured network of nodes; information populates the network from decentralized information sources on a push or pull basis; it is possible to insert information analysis agents, to see what's new, to add information, and to simulate what-if scenarios. This sort of situation assessment can be used for military purposes, by corporations tracking the market, or by communities, e.g., JTF/ATD or a Guide to Dallas.

Fabric of Knowledge: focuses more on content and how to structure and elicit it and where to locate it so as to enable situation assessment in DoD and in everyday life.

Data dissemination: The soldier in the field needs fisheye views of information: specified quality of service, just-in-time updates, dynamic replication, and customized service, under conditions of limited and varying network bandwidth, changing connectivity, arbitrary delays, and variable and dynamic transmissions are all needed to provide just the information needed in the form needed just when it is needed and to accommodate different latencies and platform characteristics. Sometimes only unexpected information needs to be transmitted [Davidson]. Information has time value, and cost-benefit profiling will be necessary to help determine whether it is worthwhile to send the information, given timing considerations (the time of acquisition of the information, the times of the events described by the information, and the time required to transmit it), the cost (in terms of bandwidth and other resources) of sending the information, and the importance of the information.

Data filtering: The commander needs information filters, negotiation, analysis, situation assessment, all from distributed sources, constantly updated, with alerters for what is new, and "interestingness" metrics.

DoD JWID 95 included several scenarios. One was Tempo Brave, a disaster relief scenario in Asia (two countries, one with earthquake, other with typhoon). In general, the services want to be able to support multiple major military and/or civilian crises simultaneously. For JWID 96, JTF/ATD are planning for more collaboration beyond Videoconferencing and Whiteboards, continuous replanning, visualization, more of the impression of an integrated system, more robustness (server replication and recovery, user visibility into communications problems, and the ability to work over degraded networks.

Digital Libraries

The goals of theDigital Libraries program are locating, organizing, filtering, summarizing, analyzing, protecting, and updating information of any kind to make information dissemination and information fusion easier.

Agile Manufacturing

There are many aspects to agile manufacturing, which is the general ability to flexibly manufacture large or small, standard or custom parts and systems "better, faster, and cheaper." Some sub disciplines:

concurrent engineering - The goal of concurrent engineering is to reduce design cycle time. Concurrent engineering is the ability to concurrently design various views and partitions of a product while insuring that cross-view integrity constraints are met; it also provides the ability to link design and manufacturing. A key assumption is that the design is distributed, often across organizational boundaries.
supply chain scheduling - The objective here is to reduce manufacturing cycle time by scheduling and coordinating a collection of designers and manufacturers to complete products more quickly.
standards for domain modeling - The PDES/STEP community is defining standard representations using the modeling language EXPRESS and mapping these to OMG IDL. To the extent that manufacturers describe parts via STEP domain data standards, these representations are easier to exchange among producers and suppliers.

Several other communities beyond the manufacturing community are beginning to work on standard domain libraries: The OMG Manufacturing, Financial, and Healthcare Task Forces are examples as is the Open Geographic Information Systems Consortium.

Technology Drivers

PC as Personal Connectivity

Fixed or mobile, the standalone desktop metaphor for a personal computer is quickly changing to personal connectivity. End-users at home, in business, or in DoD need personal views of the web conditioned on their personal working environment and their roles, consisting of environment settings for location, time, machine, network, mood, need, e.g., local/movies where local = Dallas near here via GPS.

Network as Global Computer

The protocols that bind the Internet together synthesize a global computer out of a large collection of processors and networks. How do we program it? What is its model of computation? Telnet reduces the Internet to a multiprocessor while FTP reduces it to a file system. CORBA might reduce it to a distributed object system and, with scripting languages added, to a distributed agent system. However, the Internet has its own particular characteristics which need to be taken into account in any attempt to impose more familiar programming models on it [Cardelli].

Objects for the Masses - Populating the Global Object Base - Uniform Information Space

Some have characterized the WWW as the network industry's first attempt at building the world's most primitive distributed object-oriented database system.

If we can encapsulate files as objects in an upward-compatible way so that files are strongly typed, there is support for inheritance, and there is support for methods, then we could automatically populate the Global Object Base. If it is just a matter of implementation, then files, RDBMSs, and OODBMSs may be similarly encapsulated so there is not a within-between distinction with respect to the content of a file and the structure of the file system and so that every DBMS object can be named (see, e.g., [Liljeberg, Heln, & Raatikainen], [Bowman & John] ). Also, see Semantic File Systems.

Mobile Computing and Transportable Object Computing

With virtual machine technology, applications can be built on a single platform and can run on almost any other platform. Objects can be transported across machine and network boundaries and can communicate with other objects via transports such as IIOP and DCOM, without considerations of diverse underlying platforms or operating systems

Web delivers global X (where X = hypertext, apps, distributed transactions, real-time multimedia)

The Web now delivers global hypertext. Now, can the next generation become:

a globally distributed application delivery system?
a globally distributed transaction processing system?
a globally distributed real-time multimedia content delivery system ?
a globally distributed storage system?

Frameworks and Componentware

Architectures such as OMG's potentially provide a number of architectural properties like seamlessness, safety, evolvability, security, and scaleability. Componentware is a path to the future of component specifications and reusable implementations. Federating object services is the way to scale OMG to the Internet. Monolithic DBMS and stovepipe systems are bad but pervasive. Wrapping of Internet Services and legacy information systems fits here. Standards are needed but so is customizability for DoD and industrial applications. Widely available reference architectures and reference implementations are needed.

Business Object Facility

Business needs a large collection of common reusable standard component parts that can be delivered to workstations and assembled automatically into applications. Business objects and business rules need to be understandable by end-users (as opposed to ordinary code understandable only by application developers). Business wants services and products, not just simple applet animation. Reuse of Perl scripts and Java applets is hard due to tight application coupling, opaque semantics, and lack of abstraction [Carlson].

Document Management, Collaborative Authoring, Active Documents

Firms increasingly provide product specifications and technical support to their sales representatives and customers via the network. They need interoperable multi-vendor configuration support; and must be able to catalog, index, search, and view secure active compound documents, in some cases applying restrictive terms to the activity (such as requiring payment) which may involve negotiation.

Today, we generally either pass authorship around to complete a compound multi-authored document or we send our sections by email to a designated editor. We are constantly regenerating the document in different formats and doing conversions, some manually, some semi-automated to compress or otherwise transfer the document between environments. The many conversion programs now available help. But it would be nicer and more seamless if, while reading a document, we could check out sections, annotate them with our comments or revise them, scan old versions, possibly issue structured queries, be notified if others made changes, and more. And do this over the net, with different parts of the document stored remotely, with appropriate concurrency control and security support, and with most data conversions automated. This scenario is more interesting if the kinds of data being shared also include CAD designs, patient records, digital libraries, and all those legacy databases out there. It is hard to believe that all this can be done without objects, for instance, using only Netscape helpers and typeless URLs. It is easy to see how an object layer would be useful. Sophisticated compound document architectures such as OpenDoc are aimed at a vision of "composing applications" as one now composes documents, and provide glue mechanisms for connecting heterogeneous applications.

This research is sponsored by the Defense Advanced Research Projects Agency and managed by the U.S. Army Research Laboratory under contract DAAL01-95-C-0112. The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied of the Defense Advanced Research Projects Agency, U.S. Army Research Laboratory, or the United States Government.

© Copyright 1996 Object Services and Consulting, Inc. Permission is granted to copy this document provided this copyright statement is retained in all copies. Disclaimer: OBJS does not warrant the accuracy or completeness of the information on this page.

This page was written by Craig Thompson and Frank Manola. Send questions and comments about this page to thompson@objs.com or fmanola@objs.com.

Last updated: 1/3/97 sjf

Back to Internet Tool Survey -- Back to OBJS