The use of the national information infrastructure and high performance computers in industry

We divide potential NII (National Information Infrastructure) services into five broad areas: Collaboration and televirtuality; InfoVISiON (Information, Video, Imagery, and Simulation on Demand), and digital libraries; commerce; metacomputing; WebTop productivity services. The latter denotes the broad suite of tools we expect to be offered on the Web in a general environment we term WebWindous. We review current and future World Wide Web technologies, which could underlie these services. In particular we suggest an integration framework WebWork for high performance (parallel and distributed) computing and the NII. We point out that pervasive WebWork and WebWindows technologies will enable, facilitate and substantially accelerate such complex software processes on the NII. We briefly analyze seven broad application areas: society; business enterprises; health care; defense command and control, and crisis management; education; collaboratory; manufacturing. We contrast their use of NII services with a more detailed examination of the manufacture of complex systems, such as aircraft and automobiles. This application stresses the NII but there is a remarkable opportunity to develop new manufacturing practices that offer cost savings and reduced time to market.


Introduction
We are at a remarkable time in high-performance computing and communications (HPCC) technology.Research and development over the last decade or so is maturing rapidly, and can be applied in several important ways to real world problems.Further, most progress is perhaps from the con uence of technologies and indeed academic disciplines, including parallel and distributing computing, networking, and information science.In Section 2, we review the current situation with current World Wide Web (WWW), National Information Infrastructure (NII), and HPCC.We set the standard multi-use NII vision with infrastructure driven initially by the entertainment and information industries, but applied broadly over a range of applications from education to manufacturing.In the speculative Section 3, we suggest an emerging World Wide Metasystem, where the user interacts with WebWindows|the operating system of the WWW.We describe how a network of combined compute and Web servers can be harnessed in WebWork as a distributed or parallel compute engine.We believe that Java, from Sun Microsystems, is (an example of) critical technology, and we illustrate its use in some examples, including data ow computing, software engineering, and education.We discuss other critical emerging Web technologies, including the integration of relational and distributed database technology, and the unifying middleware, WebScript.
In the nal section, Section 4, we tie the technologies to a set of base multi-use NII services whose use is illustrated by seven application areas, including manufacturing, which is discussed in a little more detail.2 The WWW, NII, and HPCC today 2.1 The Current World Wide Web Technologies Table 1 and Figure 1 summarize some important aspects of today's WWW 2], 3], and 4].Clients have sophisticated display capabilities, and request and receive hyperlinked material from servers.Documents are typically written in HTML, which is a simple subset of SGML, which supports hyperlinked information via URLs 5]|universal resource locators|specifying location of related material.HTML is a dynamic standard and extensive Web Search engines, such as YAHOO, HARVEST, WAIS, LYCOS|early distributed database access and important new functionality is being added 6].Servers and client use MIME as the data format, and this can be thought of as an extended email syntax with headers specifying format or method (program to be executed by remote procedure call, RPC) followed by the body of message, which is typically desired data.The Common Gateway Interface (CGI) is a standard way to add functionality to WebServers with software that typically uses PERL as a convenient rapid prototyping environment o ering good access to system resources, les, and document manipulation.

Further Critical Technology Building Blocks for the NII
We envisage an architecture for the NII, shown in Figure 2 that builds applications on top of general (multi-use) services, which themselves ride on a collection of technologies, and middleware, which we term generically WebWindows, and discuss in Section 3.
In Table 2, we list some important pervasive technologies that will be critical building blocks of the future NII.This includes the basic hardware architecture where there is general agreement on the importance of ATM.Currently, ISDN is being deployed quite widely, and it will be important to see if this performance (128 Kbits/second) is su cient to enable many NII applications.One to two megabits/second performance is needed for NTSC quality digital video with MPEG compression.However, either better compression (such as Wavelet) or quality reduction (smaller images) is su cient for many applications, such as education (where small clips|not two hour movies are needed) and collaboration.The latter illustrates a typical NII/Web issue.There are several excellent video conferencing systems, such as Intel's Proshare for the PC.However, a good collaboration environment requires not just video exchange of the participants, but also common whiteboards and other text and graphics information exchange.The latter are clearly best implemented with Web technology, and I see that the days of stand alone collaboration systems are numbered|video conferencing will be integrated with the Web and then will evolve to a full televirtual environment embodying such technologies as VRML and MOO's|the most elegant multi-user virtual worlds.Other integration challenges are seen in the message passing arena where we should reconcile parallel processing technologies, such as MPI and PVM with MIME and HTTP of the WWW.

An Entertainment and Telecommunication View of the Early NII
The NII will on one hand build on the Internet and the WWW.On the other hand, much of the infrastructure will be motivated by some early business opportunities, which roughly can be thought as digital versions of today's cable TV.As shown in Figure 3, several industry sectors are teaming and competing for this market, which also includes interactive shopping and the digital equivalent of the current video rental business.
In Figure 4, we try to quantify this vision assuming that some 100 million clients would be connected to the digital media information system.The year in which this becomes reality depends a lot on both technology and government|some regulations could sti e competition and slow digital deployment; others would speed the process through encouragement of provision of digital services to remote areas that might be uneconomic in a purely market driven process.The NII will link hybrid delivery technologies (including ISDN, cable, ADSL, wireless, mobile systems) for the o -ramps to ATM and satellite, which can be used for both o -ramps and the long distance trunks.Further, we expect large-scale Massively Parallel Machines (MPPs) to be part of a network of video servers, and more generally, WebServers that supply information to these clients.We term this scenario InfoVISiON for Information, Video, Imagery, and Simulation on Demand.This includes the storage, query, and dissemination of this wide range of multimedia data.There is surely at least 100,000 hours of interesting video material in the archives of Hollywood studios, CNN, Routers, and network TV.If compressed in MPEG format, this corresponds to some 100 Terabytes of needed storage capacity.MPEG2 or other format, such as motion JPEG, would require much more storage and are probably necessary to support editing and other video production applications.Note that the current WWW has only a few percent of its storage devoted to video| the future NII will be dominated by video data.Each NII o ramp will, as shown in Figure 5, connect homes (o ces, school desks) at the rate of 1{20 Megabits/sec to the set of NII InfoVISiON servers.This rate covers the range from compressed VHS to HDTV picture quality.Note that this performance is 100{1,000 times greater than today's conventional 14.4 Kbaud  Returning to Figure 4, we estimate comparable investment in clients and servers and nd an InfoVISiON or WebServer market that is at least an order of magnitude larger than that for supercomputers.This illustrates that it is the NII, and not large-scale number crunching, which is the best opportunity for parallel processing.Notice also that the compute power contained in this future NII is some 10{100 PetaFLOPS|far larger than the compute capability of an individual TeraFLOPS supercomputer.This motivates our interest in WebWork, described in Section 3, and aimed at extending Web to compute servers, and harnessing the power of the World Wide Metacomputer.
One could view this InfoVISiON scenario as the most amazing clientserver application with 10 8 clients and 10 4 large servers.However, Vice President Gore has articulated NII democracy with everybody able to host Further new technology, just as Java and WebTools (Section 3), suggests instead a heterogeneous server-server architecture with 10 8 distributed nodes.This can be viewed as a fascinating parallel computer with many more nodes than traditional tightly coupled systems (by a factor of 10 5 compared to typical large 1,000 node MPPs).It appears that although powerful, the communication backbone will not allow every \client" long distance simultaneous access to every other client or server.Rather, we must enforce the guiding principle of all computer architectures|namely, data locality.This is illustrated in Figure 6.When Jurassic Park VI is released on the Hollywood Server, one will not have everybody accessing it there at an average distance of some 1,500 miles.Rather, this \hot" movie will be cascaded down through the hierarchy of servers so that any individual will nd it on a server a few miles away.This strategy reduces the needed ber for the NII trunks by a factor of about 100.
Figure 7 illustrates a related point.We consider a WebServer as a system implementing the software model of Figures 1 and 2. This could be a server with full input and output access to the World Wide Information, or it could be aimed at a smaller organization with rewalls, as necessary,  to enforce security.We see that the mix of distributed and conventional (PC, Mainframe) software of Tables 1, 2, and 3 makes WebServers the most attractive technology to build a business enterprise support system.
3 Features of the Future WWW and Implications for HPCC

WebWindows and WebTools
We assume that current operating systems for individual computers|such as UNIX, Windows95, Apple Macintosh|will have less signi cance in the future.Rather, we will use the open non-proprietary WebWindows, which is the operating environment of the Web built from the collection of current and so this environment can manage systems at all levels of granularity| from individual machines to the full Web metacomputer.We can illustrate the meaning of this concept with a typical productivity tool like Microsoft Word.Currently, one develops a separate version of Word for each machine|PC or Macintosh|and lament its unavailability for UNIX.In the future, one will develop a single product WebWord, which uses a suitable combination of client and server Web technology so that the machine dependence is isolated in Web servers and clients, and applications are architecture independent so that through a universal browser, they present the same interface on all machines.Wojtek Furmanski, at NPAC, has illustrated the WebWindows vision with a set of CGI programs, WebTools that implement such universal tools for le management (create, delete, copy) and navigation|roughly equivalent to nder on the Macintosh or program manager on the PC; a sophisticated Webmail built on top of mh, the UNIX mail system; and HTML editor.
As shown in Table 3, we see that there will be a complete set of \Webtop" productivity tools supplanting and extending functionality of Word, Excel, Persuasion and similar tools.These will join capabilities indicated in Tables 1 and 2, and further applications to create a WebWindows environment that cannot be matched on any individual computer running conventional operating systems.WebWindows draws its power partly from the distributed computing simulation and information processing power| as recorded in Figure 4; more important is the creative energy of the \Web software engineers" (hackers) community interacting in the design and implementatio of WebWindows.Critical to success is open standards for interfaces and protocols|quality proprietary modules but with open interfaces are quite consistent with continued success of the Web.It is interesting to contrast the dynamic organization of the WebWindows development \team" with the more traditional static hierarchical structure used by IBM, Digital or Microsoft in the development of their major operating systems.

WebWork|A Pervasive Technology Base for HPCC
In many ways the HPCC program has been a great success and demonstrated as we like to say that \Parallel Computing Works!" 13].However, MPPs remain rather di cult to use.This partly re ects inevitable problems in parallel programming, but also the immaturity of current MPP software development environments.We know how to do build much more powerful MPP software systems than are currently available.This is not surprising, as parallel (and distributed) computers are clearly the most complex computer systems.Thus, one would expect that it would require more e ort to build a software environment for an MPP than for PC.Unfortunately, as shown schematically in Figure 8, the PC market is about two orders of magnitude larger than that for MPPs.Correspondingly, the software environment for PCs (and to a lesser extent) workstations must be much better than for MPPs.The situation can only improve if the size of the MPP market increases dramatically, and this currently appears unlikely in the technical computing area.We notice that the WWW is, as described in Section 2, a distributed computing environment associated with pervasive technology base and a corresponding large and vital software development infrastructure.In WebWork, we suggest, in collaboration with Boston University and Cooperating Systems, building a parallel programming environment on top of \Web Technology."This is not a terribly well de ned statement as it implies and assumes an extension of the Web in many ways from information servers to combined compute-information servers.These will feature many enhancements from today, including integrated security and low latency HTTP-NG protocols connecting multi-threaded Web servers 14].Much of this is discussed in 15]. Figure 8 shows a traditional HPCC strategy of porting MPP technologies so they can run on more available technologies, such as networks of PCs or workstations.WebWork takes the opposite approach of extrapolating from the base to the tip of the pyramid.The classic HPCC approach has the problem that it does not naturally produce technology used by and hence supported by a broad community.WebWork has a potentially equally serious problem that the extrapolation might \miss the top of the pyramid" i.e., produce a system that either did not meet the needs of parallel computing and/or produced very ine cient parallel code.We believe that careful design can avoid this problem.For instance, we are leading a group developing an open parallel compiler runtime that will support HPF (High Performance Fortran), and parallel C++ on several platforms 16], 17].This runtime embodies key parallel computing synchronization and collective communication and computation primitives.As shown in Figure 9, Webwork will reuse such software, but provide an attractive front end and use suitable low latency Web compatible message passing systems.This concept when re ned with a careful mix of interpretative and compiled environments leads to WebHPL|a general parallel language combining the lessons of expressing and implementing parallelism Often, we have worried about the concept of parallel software engineering (PSE), but found no convincing approach.Now, I realize why I had diculties.Good software engineering requires a good support (productivity) environment, and this is absent from all current parallel systems.WebWork has a clear mechanism for PSE by using the natural linkage of information and computing in the extend Web|this we term the Virtual Software Laboratory.We will illustrate this in Section 3.4 with our WebFlow concept implementing compute, information, and project management in a uniform data ow framework.

Some Emerging Web Concepts and Technologies
In the near term, many new innovative technologies will be critical in building WebWindows as well as services and applications on top of it.These  Java 18]|This is a very powerful interpreted C++ like language (script), which, as in Figure 1, can be used to build fully interactive clients that allow one to build balanced client-server (server-server) systems.The CGI mechanism, as used in WebTools, is powerful but has limitations as extra functionality can only be added at the server.
VRML 19], 20] can be thought of as a three-dimensional HTML that allows universal description of physical objects and graphics actions on them.There is some confusion as to relative importance and role of VRML and Java, but roughly one can consider VRML as the future data structure and Java the future language of the Web.VRML will be useful in such areas as telecollaboration, multiplayer gaming and distributed manufacturing, because it allows the interchange of virtual worlds and commercial product designs.
PERL5 is a useful extension of PERL4 with full object oriented characteristics and extended pointer (array) constructs.Again, there is some unclear competition with Java, but the languages are optimized in di erent ways.Java is partially compiled and will produce code that executes faster than that from PERL(5).PERL5 is better suited for rapid prototyping and further has excellent special capabilities built in for dealing with operating system functions and text processing.Performance of future combined Web-Compute Servers will be enhanced by good light weight multi-threaded systems combined with new protocols, such as HTTP-NG.These technologies are linked together into loosely coupled integrating concepts summarized in Table 3.We have already described WebTools as an early prototype of WebWindows illustrating primitive Web operating system services.Again, WebTop publishing and productivity underlied our discussion of software engineering in WebWork.
As already mentioned in Section 2.2, we see Web database systems Web-DMBS combining distributed (as in Harvest, Lycos, etc.), object (as in VRML) and relational (as an Oracle) capabilities.In Figure 10 Multi-user shared distributed virtual worlds of information and simulation.
WebSpace denotes this combination leading to a full televirtual Web collaboration environment.
An important deeper and perhaps more controversial concept is Web-Script.This denotes the complex NII middleware of scripted languages where we expect no universal solution but a loose federation where each component has di erent optimizations|VRML for three-dimensional objects, PERL5 for text, Telescript for agent based communication, Java, and MOVIE 22] for computation, etc.

WebFlow|A Simple Web Data ow Interface
We consider WebWork as having three layers World Wide Virtual Machine|the basic network of Web-compute servers supporting MPI, HTTP, MIME, etc. data transport protocols and format.The intermediate integration layer, WebScript.At the highest level, we have many domain speci c user interfaces optimized for di erent purposes.There are, of course, Web browsers for reviewing information and WebHPL for parallel programming.There will be several other such environments, and below we describe one example, WebFlow.Another simple rather general high-level interface is WebFlow under development at NPAC.WebFlow is particularly interesting because the same framework can support both computing and the software engineering process that produces the software for the computation.Thus, it illustrates the power of WebWork's linkage of computing and information processing.Data ow is a very useful programming paradigm rst popularized with AVS and Khoros aimed at visualization and image processing.However, it has since been successfully applied to general coarse grain software decompositions 23], 24], 25], 26].We are building this functionality into WebWork using Java with an early prototype of the visual Computer-Web Editor shown in Figure 11.This editor allows one to place compute modules and link them together.Figure 12 illustrates an example of WebFlow application in the area of work ow management|a software project management.Here, we suppose each software developer runs a personal WebTools server implementing the hyperlink enabled CASE tools Hysource supplied with WebTools.These personal servers are connected (via the WWVM in WebWork) to agent and manager servers also shown in Figure 12.The agent (database con guration) server receives automatic noti cation from developer servers on each software volume update.The agent server uses customizable thresholds to decide when to send a report to the manager or a deadline reminder to a developer.Note that the underlying database is stored in distributed fashion on the collection of WebTool servers.

Introduction
Returning to Web software architecture of Figure 1, we now discuss services and applications.These are not clearly de ned for services are essentially generic applications, and most applications are complex metaproblems 27] built recursively from services and \sub-applications."Thus, there is a grey fuzzy line distinguishing services and applications.For this paper, we have chosen seven applications discussed in Sections 4.2 and 4.3, which we relate to the ve services described below.

WebTop Services|Publishing, Productivity, Software Engineering
This includes all the base WebWindows functionalities that we have already discussed.

InfoVISiON|Information, Video, Imagery, Simulation, ON demand
This includes base database storage, management, query, and dissemination of the full range of multimedia archives of the World's distributed digital libraries.As already discussed, we can expect hundreds of Terabytes of available information dominated by video data.Note simulation|such as access on demand to a weather model|is included in this service.
Commerce|Digital Cash, Security, Authentication, etc.This collection of services enables electronic commerce, including online banking and shopping.These services are also essential for the use of the WWW for processing and exchange of proprietary data.

Collaboration|Real-Time Interactive and \Batch"
This includes desktop video conferencing, three-dimensional graphics MOOs, geographically distributed CAVEs leading to full televirtual interactions.As discussed earlier, a wide variety of other types of interactive information exchange is necessary.This underlies the concepts of collaboratories (virtual research groups or scienti c laboratories), and the virtual company of the next century's agile manufacturing environment.In the more static mode, we see work ow and con guration control (cf. Figure 12, which allows tightly integrated projects, such as those to build a complex system such as an aircraft (see Section 4.3) or a large software module with a distributed team (Section 4.2(f)).
Metacomputing|The worldwide collection of computers organized in a subgroup as a single computational engine for simulation or information processing.This service can be used to control remote medical and scienti c instruments; search the world for information; or link computers in di erent companies for a multi-disciplinary optimization of a new vehicle.Some services listed above can be already prototyped in terms of today's Web technologies, for example base WebTop or early Collaboration services.Some other services are still waiting for their pervasive enabling technologies, such as physical infrastructure that will enable InfoVISiON or security that will enable Internet Commerce.Finally, the computationally extensive NII services, characaterized above broadly as \Metacomputing" require a major extension of the whole Web paradigm, currently still focused on static page services, but already gradually expanding towards computation and interactive simulation via technologies such as Java, or WebWork discussed in Section 3.4.Based on current and emergent Web technologies discussed in Section 3, we can start looking into the future to envision the coming generation of critical NII applications and high level services discussed in the following section.

Some Important Applications of the NII and Critical Services a) Health Care and Telemedicine 28]
Critical to much NII use in health care are large-scale distributed databases to store patient records and medical instrument data.This should lead to more cost-e ective health care with sophisticated database search techniques reducing fraud and waste while the large collection of uniformly prepared records will help emergency care, and the identi cation of \best of practice" care plans for particular medical problems.Telemedicine (or remote medical care) naturally uses collaboration services, as well as the metacomputer capability for control of remote instruments.This could extend to use of a full televirtual environment for virtual-reality controlled remote surgery in battle eld or accident scenarios.

b) Education 29]
The NII is expected to make major impact on education both at the K{12, and University level.Indeed, the concept of the Virtual University 30] indicates that some fundamental concepts underlying the traditional residential college may need to be rethought.The bases of the educational use of the NII is collaboration and InfoVISiON.Rich collaboration environments including three-dimensional MOOs will support student-student and student-teacher interactions around the globe, and help students in rural areas and those home sick.Info-VISiON will allow full interactive and explorative multimedia curricula that can be expected to be more successful than current analog video delivery of distance education.We are currently working on Java and WebFlow based tools to support educational delivery systems.Note that hyperlinking capabilities of the WWW are just as important as multimedia, for it enables student controlled exploration as opposed to current rather rigid model of learning embodied in a book.Geographical Information Systems will allow virtual eld trips using the NII.

c) Society
The initial motivation for the installation of commercial NII communication instrastructure is digital delivery of current CATV and video rental services.This InfoVISiON capability will naturally generalize to a full range of multimedia products on demand, but it is unclear what novel services will be pro

e) Defence (Command and Control) and Crisis Management
A critical feature of this class of application is the need to make quick decisions in the presence of incomplete inaccurate data.We can term this scenario as judgment (as opposed to decision) support with the less precise context leading to the need for an InfoVISiON exploratorium.This will support multiple, but related video streams (say from di erent reporters of a given event) that need to be linked in time.
Critical features of the network used to support crises is that it links a real-time, \come as you are" adaptive mix of computers and people.This is a complex metasystem mixing dynamic metacomputers and collaboration.
There are major needs to search community and government databases for immediate information on resource location and availability.As these databases are likely to be unfamiliar to judgment makers, metadata summaries in common formats are essential.
Collaboration will be required to link commanders in the eld, specialized anchor disks (domain experts) and those in the eld.A special issue will be the importance of excellent user interfaces as users will be tired and searching unfamiliar data.Geographical information systems will be very helpful as judgments will depend critically on spatial data.

f) Collaboratory|The Virtual Science or Engineering Laboratory
As desired by Wulf 31], the NII and collaboration services could enable much richer forms of scienti c and engineering collaboration.

Use of the NII in the Manufacture of Complex Systems
We describe the application of the NII to the manufacture of aircraft, automobiles, and similar complex systems in rather more detail than the previous examples.This analysis stems from a NASA sponsored analysis of the NII requirements for a future concurrent engineering concept called ASOP (A ordable Systems Optimization Process).This involves an industry consortium MADIC (Multidisciplinary Analysis and Design Industrial Consortium) with a team from Rockwell, Northrop, Grumman Vought, McDonnell Douglas, General Electric, and General Motors.Interesting parameters specifying the scope of the design of the next major aircraft include: Construction will be led by a consortium of some six major companies and 20,000 smaller subcontractors.
The number of engineers involved could be about: { 50 at conceptual design { 200 at preliminary design { 2,000 in nal design { up to 10,000 for manufacturing and development ASOP involves multidisciplinary (multicomponent) optimizations (MDO) involving 10,000 separate programs that would be run in linked clusters (e.g., 10 programs at one time) for a set of speci c design decision optimizations.These programs vary over a wide range from the full air ow simulation around a plane to a simple expert system to plan the best location of an inspection port to minimize maintenance life cycle costs.Such a manufacturing enterprise is an exciting and demanding challenge for the NII.First, note that the NII and its associated services is e ectively essential for this application because the expertise and infrastructure needed for the design and manufacture of new aircraft is spread geographically through the country and perhaps globally.This expertise needs to be linked (by NII) to perform collaborative and coordinated design and simulation.If the NII did not exist, equivalent capabilities would need to be supplied by the involved companies, and indeed this has happened, using private lines as infrastructure, on some earlier projects.What are some key NII services needed by ASOP?Compared to the previous NII application's discussed in Section 4.2, manufacturing requires a close and integrated coupling of the very many people and computers involved.We are linking them in the design and manufacture of a very precise entity as opposed to the looser coupling required by, say, collaborative scienti c research.Remote surgery in Section 4.2(a) is an example of another such close integration requirement for the NII.Metacomputing and distributed database support will have strong requirements to support the large number of linked programs connected to a logically central, but physically distributed design database.Work ow support must include con guration management and strong coordination with structured updates of the design database.
Standards and security will be needed to link people and software from di erent organizations.In particular, e cient security for large les will be needed.
Clearly NII and novel optimization techniques must present a good evolutionary path to allows re-use and incremental upgrading of exist-ing software and people infrastructure.This implies good \wrapper" technology to support use of existing software modules with new interfaces.
Finally, the NII collaborative services will be stressed by the close coordination of the large number of engineers needed in design process.We can use this application to brie y discuss the use of parallel computing in industrial simulations 32], 27], and 33].This has not been as successful as many people hoped.For ASOP, we can see some aspects of the problem.
Parallel processing can certainly be of value in simulations for propulsion, aerodynamic and probably other areas.However, these are a small fraction of the tasks (remember we mentioned 10,000 programs in ASOP) needed to design a new aircraft.Thus, we nd a variant of Amdahl's law|parallel processing can e ectively reduce needed computational uid dynamics (CFD) simulation time.However, if this is only a fraction of x (x 0:1 perhaps) of total endeavor, we have a speedup of at best 1 + x 1:1!As the aerospace industry adjusts to reduced DoD spending, and the construction of fewer military aircraft, it is hard for the industry to invest in new technologies with an unclear return on investment.
Thus, we see parallel processing on its own, insu cient to develop approaches to manufacturing.Rather, we need the integration of high speed networks and computers envisaged by the NII.Further, we require several basic NII services|security, metacomputing, collaboration, wrappers and agents, work ow with con guration management|to be well developed.Thus, we can expect the NII and, in particular, parallel processing to have a profound compact on manufacturing that will be great value to the National Enterprise.However, this is not an easy or short task, and we can expect signi cant government investment to be needed in basic precompetitive technologies and services.Industry is not likely to be able to make the necessary long term investments on its own.Correspondingly parallel computing will be used in a major fashion in manufacturing, but not in the near future, and not without continued thoughtful investment by industry, government, and academia.

Summary
We reviewed above the most plausible near term NII services and applications.Our claim is that this vision can be accomplished most realistically in the framework of pervasive, truly open Web technologies such as WebWork and WebWindows discussed previously in Section 3. Indeed, the traditional corporate computing failed to establish even the most elementary software standard, namely the cross-platform desktop publishing system.In contrast, Web already enjoyed a smashing success in stabilizing the base electronic information service|a homepage|and continues the process of developing and disseminating more advanced computing standards such as Java based graphics user interface etc. New Web technologies fully utilize and build upon already established standards, for example NPAC develops WebTools using WebTools CASE tools, or in a similar way Sun delivers Java together with complete on-line HTML documentation for all object classes.We expect the anticipated Web technologies discussed here such as WebWork and WebWindows to continue this bootstrap process and result in software environments of unprecendented power, functionality and quality, and yet fully open and pervasive.WebWork standards for module encapsulation and intermodular communication will enable complex manufacturing processes such as required by ASOP and will e ectively turn our vision into the NII reality over the period of the next 5{10 years.

Figure 1 :
Figure 1: A simple representation of the software architecture of the WWW showing emerging advanced services, as well as the current technology

Figure 4 :
Figure 4: An estimate of the communication bandwidth and compute capability contained in the NII and supercomputer industries

Figure 5 :
Figure 5: The basic InfoVISiON scenario as seen by a home in the year 2000 with an intelligent settop box interfacing the digital home to a hierarchical network of InfoVISiON servers

Figure 7 :
Figure 7: A Web server can be used at any level of organization from an individual home to the entire universe PCRC embodies the Parallel Computing Synchronization and collective parallel algorithms and runtime that will enable efficient Web-based computingReplace user interface of HPF or HPC++ with the Web(Work) and use pervasive Web Technologies in infrastructure (World Wide Virtual Machine -WWVM)

Figure 9 :
Figure 9: Integration of HPCC technologies with the Web illustrated for the parallel compiler runtime

Figure 10 :
Figure 10: The three axes of multimedia extensions to give interactive services with di erent characteristics

Figure 11 :
Figure 11: Early prototype of the Java-based WebFlow Compute-Web editor allowing interactive placement and linking of software modules

Table 1 :
Some early WorldWide Web (WWW) technologies and services Clients (such as Mosaic and Netscape) support browsing of hyperlinked documents, but have no internal interactive/compute capability 7] Servers read HTTP and deliver requested service to client HTML|a document format supporting hyperlinks HTTP|a Transport Protocol de ning Interaction between

Table 2 :
Some important real world technologies that are (to be) incorporated into the WWW.

Hierarchical Server Network The world: Global Information Infrastructure New York State Server Shopping Malls Corporations Universities School Districts National Master Server (Hollywood for Movies)
Figure 6: A typical hierarchical server network depicted for a master system in Hollywood cascading down with a fragment of node systems shown for central New York

Table 3 :
Some emerging high-level WebWindows integration concepts

HPCC needs a large enough market to sustain technology (systems and software) This implies that we look at both Grand Challenges and National Challenges, but we suggest this is not enough: WebWork Builds HPCC technologies on a broad not niche base starting at bottom (Web, PCs) not top (MPPs, Supercomputers) of computing pyramid
table.The NII will also enable advanced community networks that will given an interesting new view of local society, and could have a very important impact on local government.Commerce services will enable shopping and banking on the NII.In-foVISiON services would include multimedia digital yellow pages and online catalogs.The yellow pages could use an NII implementation of geographical information systems while virtual reality interfaces could allow you to try out new products such as cars and clothes.Collaboration services would support gossip as in today's computer forums and generalize the popular talk shows on radio and TV.