eBusiness: The Hope, the Hype, the Power, the Pain

(© Jack M. Wilson, 1999, 2000)

The Technologies of eBusiness

God may be subtle, but he is not malicious – Albert Einstein, Physicist

It's kind of funny that Moore's Law is what I'm best known for; it was just a relatively simple observation. –Gordon Moore, Founder and former CEO, Intel

Computing, communication, and cognition are driving the fundamental changes in the way we live, love, labor, and learn. The advances in computing are obvious to most, but the growth of communications technologies has probably been more dramatic. The eBusiness revolution is fundamentally about human interactions. Psychology and the cognitive sciences have illuminated the ways that people interact and learn together. How does a business learn about its customers? How does a consumer learn about new products? How does a manufacturer or developer support customers? These are all timeless questions, but technology has changed all the answers and will continue to do so for the next decade or more. What can we know with great certainty about the future of technology and what can we only guess at? There have been a number of efforts at projecting the future.[1],[2]

To understand the technological changes that underlie the opportunities in eBusiness, one must include a backward look at the technologies that have brought us to this point and then take a prospective look at the basic sciences that will create new technologies that will continue to drive further change. Some things can be predicted to be as relentless as death and taxes. Others are as unpredictable as the weather. Understanding which is which and developing strategies for dealing with relentless change and uncertain results is a key component of eBusiness and the management of information technology.

Moore’s Law

Gordon Moore first observed that computer power was doubling every two years. That was later revised to 18 months. This means that since the invention of the microprocessor, the performance of a microprocessor has been doubling every eighteen months. This is directly related to the number of components that can be fit on a chip and is thus related to the minimum size of each component. Advances in basic physics and engineering have kept this law accurate for over five decades! The basic physics is in place to keep this going for another few doubling periods at least. Moore's Law is the simplified statement that computing power is doubling every 18 months. An alternate perspective states that the cost of equivalent computing power halves every 18 months. Either way, it creates a real challenge for everyone to keep up. On the other hand, it provides an extremely reliable roadmap to what technology platforms will be available in the future. This and the relentless doubling of bandwidth are the two most predictable events driving information management.

How long can this last? Gordon Moore thinks that it might begin to break down about 2017, but other observers are confident that physicists and engineers can come up with the new devices that can extend Moore’s Law for several more decades.[3]^,[4]

Bandwidth Scaling Law

For the last two decades, the bandwidth of communication on optical fibers is increasing exponentially just as is the power of the microprocessor. Examination of the graph below reveals two numbers of intense interest. First the optical bandwidth for commercial networks has doubled over 9 times in the eighteen year period from 1979 to 1997. This indicates a doubling time of approximately two years. Notice also that the displacement between the best bandwidth in the laboratory and the best bandwidth in commercial deployment has a delay of just about 6 years. This means that commercial deployment of bandwidth will surely continue to double for the next six years leading to an overall increase in bandwidth of 8 times! Those who work in optical networking are confident that improvements in the physics and engineering will continue to improve the best laboratory bandwidth over this period. Thus we can say with a high degree of confidence that the bandwidth doubling will continue of the foreseeable future! This relentless march of bandwidth led George Gilder to declare that the future was a future of a "Bandwidth of Plenty."[5]

Metcalf’s Law

Metcalf’s Law tells us that the value of a network is proportional to the square of the number of persons connected to that network. Although this can be given a mathematical basis, it is really more of an empirical law that must be applied in a context where all other things are equal. The mathematical basis comes from the number of mathematical interactions between n persons. The n people would have n-1 other persons that they can interact with. That is everyone but himself or herself! The total number of such interactions is then given by:

Interactions = n(n-1)= n² – n

For large values of n, n² is so much larger than n that the n is just ignored.

Investors use Metcalf‘s Law as one way to assign value to a companies network. It is not the whole story, because it matters what people do once they are on that network. How long do they stay there? Some term this “stickiness.” If people stay on a network longer then the network is said to be “sticky”. The reasoning is that the longer someone stays on a network, the more probable it is that they will purchase something and return to purchase something else. Even more important is what the company’s business model is for extracting revenue from the network.

Metcalf’s Law is said to lead to “network economics” and is sometimes described by economists as “network externalities.”

In the early stages of the eBusiness Tsunami, companies often overlooked the need to develop a business model for what they might do with their network. In the rush to establish a competitive advantage, companies “bought” customers with free goods and services. Netscape pioneered this by giving away the browser. Later entrants tried giving away PC’s and even Internet connections.

There is no denying the advantage those companies like America On-line (AOL), Amazon.com, Yahoo, and eBay have established in their networks of users, but they have only been able to convert that network advantage into a business advantage by finding a business model that leads to revenues and earnings. For some that is a quest that is still not finished.

As mentioned in the introduction, the announcement of the merger of AOL and Time Warner in early 2000 provided a dramatic example of the value assigned to Metcalf’s Law or Network Economics. The agreement granted AOL 55% of the merged company and Time Warner 45%. Time Warner had 80% of the revenues of the proposed combination and included many of the United States oldest and most respected media properties. On the other hand, AOL had 53.8 million unique visitors to its network in December of 1999 while Time Warner had 12.2. This ratio of 4.4 implies that AOL network was 20 times as valuable as Time Warner’s. Considering revenues, industry position, and network value led to more value being assigned to AOL than to Time Warner.

The list of leading networks is shown as of the end of 1999, according to Media Metrix. Much can be learned about the prospects of companies from this comparison.

These large networks often lead to Communities. Although community development is closely related to Metcalf’s Law, Reed has pointed out that Metcalf’s law really relates to one-to-one interactions among n persons. He has proposed “Reed’s Law” as an alternative.[6] If one calculates the number of communities that can be formed from n persons, then it turns out that this is proportional to n! (n factorial- or n x n-1 x n-2 x n-3….1).[7] This is an even more powerful scaling factor than Metcalf’s law. Community development has been a very important part of the Internet. It is also a key component of internet based marketing scenarios.

Exponential Growth

The mathematics of eCommerce is largely the mathematics of exponential growth. As we look at Moore’s Law, the Bandwidth Scaling Law, the growth of the market capitalization of technology based companies, the growth of eCommerce revenues, and so many other variables, we see examples of exponential growth. Every exponential growth process has a doubling time. For Moore’s Law that is 18 months, for Bandwidth it has been about 12 months. For venture capital the doubling is about 5 years. In figure 1 for the market capitalization of technology-based companies, the doubling period is slightly over two years.

Consider the growth of the number of Internet hosts:

These are growing exponentially with a doubling time of approximately 18 months. This exponential growth has driven the eBusiness revolution in a powerful and relentless fashion. Similar curves can be seen for the growth of the number of Personal Computers or the number of wired telephone lines.

In mathematics, exponential growth processes go on indefinitely with values driving toward infinity. In the physical world, exponential growth processes always end. They can never continue indefinitely since they run out of resources, people, potential markets, interest, energy, or something else. There are three general ways that exponential processes can end:

1. Saturation

2. Peaking

3. Catastrophic Collapse

When exponential processes end in saturation, the curve begins to flatten and then levels off at some higher level. When peaking occurs, the curve first flattens and then begins to fall. Catastrophic collapse occurs when the function suddenly (discontinuously) drops to some undefined value rather than moving smoothly through a curve as in the other cases.

Real world processes are never as simple and clean as the mathematics used to describe them, but these three mathematical processes describe the general fate that awaits any exponential process. Exponential growth in any system will eventually saturate, peak, collapse, or a combination of these in sequence. The only thing that is sure is that exponential growth cannot continue indefinitely. It is not always easy to determine either when or how exponential growth ends. But is always does. That is one of the many characteristics that makes investing and business so interesting.

In the case of Moore’s Law and the Bandwidth Law, considerable evidence points to the continuation of exponential growth for several more decades. It would be quite surprising if the exponential growth of Venture capital and the exponential growth of eCommerce continued for that long a time. They will be expected to level or peak. In fact the exponential curves for venture capital and market capitalization have shown behaviors that look to a combination of peaking and collapse since 2000.

Software Technologies

Network technologies and the Internet were created in their earliest incarnations for national security reasons. There needed to be a robust way to move data from site to site that could not be disrupted by nuclear war. The first networks were quickly put to work by the scientists working in the national laboratories to exchange scientific information. It was not long until the use of the network spread to scientists outside the laboratories, first to those connected through Arpanet and then, with the support of the National Science Foundation, through the internet. The network became far more useful once Timothy Berners-Lee created HTML as a way of tagging documents for simple sharing across the network. Still the network might have remained the province of the scientists if Mark Andreessen and Eric Bina, working as students at the University of Illinois, National Center for Super-Computing Applications (NCSA) had not created Mosaic. Mosaic made the Internet, the web, and HTML accessible to anyone with a computer running a simple browser.

Understanding the physical and mathematical technologies that have enabled the platform for information technology is more straightforward than understanding the software technologies. Progress in software has always lagged progress in hardware and is characterized by more visible journeys down blind alleys and sudden shifts in approach. The introduction of HTML by Tim Berners-Lee, a physicist at CERN, in 1991 [http://www.techreview.com/articles/july96/bernerslee.html] was done to enable communication and dissemination of information among scientists. The creation of Mosaic by Marc Andreessen and Eric Bina, students at the University of Illinois opened the Web to the rest of the population. When the limitations of HTML began to chafe, Bill Joy, at Sun, adapted an idea from computing’s past and created Java. Sun, Netscape, Oracle, IBM, and other established companies put billions behind the new approach, but Microsoft co-opted much of the strategy and the resulting dispute and legal wrangling (coupled with the limitations and immaturity of the language) clouded the picture enough to send everyone on a search for alternative strategies.

Netscape had originally created its own approach to delivering interactivity within a browser, and after partnering with Sun dubbed the language JavaScript. Java and JavaScript had very different roots and different approaches, but at least the names sounded similar!

Microsoft both embraced Java and then began to co-opt Java into the Microsoft only environment. First Microsoft created Active-X components that were like Java in that they were downloaded from servers in live time and then ran on the client, but the Active-X components were Microsoft platform specific and could be written in a variety of Microsoft languages including the popular and easy to use Visual Basic. Active-X operated in a much looser security architecture that made it both more dangerous and more powerful that Java. Next Microsoft added Microsoft specific extensions to Java that added very desirable functionality, but broke the “write once run anywhere” paradigm of Java. In the view of many persons, there were now two Java’s – Pure Java and Microsoft Java. This was simply too much for Sun who sued and won. The resulting uncertainty caused both by the Microsoft actions and by the legal wrangling hampered further adoption of Java. The situation was clouded further when Sun withdrew Java from consideration by an international standards body in 1999.[8]

As the Java war continued between Sun and Microsoft, a new front emerged. The World Wide Web Consortium worked on extensions to HTML to create an XML standard. Microsoft raced to embrace XML, in part as an alternative to Java. In the meantime, software development in C, C++, Visual Basic, and HTML continued to dominate. Microsoft eventually introduced C# as an alternative to Java.

Throughout this process there was a creative tension between server side and client side (the end user) processing. In the old days of the mainframes, fairly dumb terminals represented the client side. The terminals knew how to render the information supplied by the mainframes (later the servers), but that was about it. In the PC era, all of the intelligence resided on the client and most PC’s were independent of the network or simply relied upon the network for file serving and printing. At least part of the success of the PC was due to the independence that this gave the user. They could be productive even when not hooked to the network, and they did not have to toady to the computing priests who lived in the glass houses.

The advent of the network based computing paradigm reactivated this old argument. How much computing should be done server side and how much computing should be done client side? Not surprisingly, suppliers of servers and enterprise wide systems liked the idea of “thin clients” accessing resources from powerful servers. Many users were suspicious that this was a return to the bad old days of centralized computing and control from the top. They also worried about the inability to be productive when not connected to the network, a worry made all the more real by the obvious unreliability of networks.

Enterprise planners however saw the advantages of potentially controlling costs and improving the return on investment in computing. They liked the creation of network tools that could be centrally purchased, centrally managed, and centrally supported.

This tension continues to the present and will likely continue for some time. Java was originally created to provide interactivity that was delivered from servers but ran on the users machine (client side processing). Today Java is often used on the server to create interactivity through delivering HTML pages to the client that the client machine simply renders. Microsoft has adapted Visual Basic to do the same kinds of things in the Microsoft environment. They have also created a server centric solution called “Active Server Pages” (ASP) that is relatively easy to implement server side and is closely integrated with other Microsoft applications.

Software versus Services

Efforts to reconcile issues of server side versus user side power eventually led to two different architectures for web based computing. One architecture was based upon Java and was supported by Sun, IBM, Oracle, and many other major enterprises. Predictably, Microsoft went its own way and created the .Net (pronounced “dot net”) architecture. Each of these architectures was designed to enable powerful web services. Many organizations felt that computing was moving into a new era in which organizations would “rent” software rather than buying it. In this model, users would pay for the use of certain software applications that were hosted, maintained, and supported by the supplier. Instead of installing new software on each computer, enterprises could simply purchase a subscription that would entitle the enterprise to access to software services provided over the network. The actually technologies that would do that could be created from combinations of the tools discussed above including: databases, JavaScript, Java, and the Microsoft .NET tools. A company that provided such applications would logically be called an “Application Service Provider” or ASP. The services provided are now often referred to as “web services.” The Gartner group defines Web services as "... a custom end-to-end application that interoperates with other commercial and custom software through a family of XML interfaces (like SOAP, UDDI and WSDL) to perform useful business functions." ASP has taken on two meanings that are quite different. The other ASP is a technology that comes from Microsoft called “Active Server Pages” that will be discussed later. This latter term refers to a server side technology that receives requests from a user’s browser and then actively creates customized pages, often from a database, for the user. It is now largely subsumed in the .NET architecture.

Databases

Today’s modern applications are dependent upon connections to massive databases and to the tools that can be used to link the data in those databases to web pages delivered to the consumer or user. True end-to-end eBusiness is absolutely dependent upon such links and upon the ability of different databases to exchange information smoothly and rapidly.

The database industry is dominated by Oracle (31.1%), IBM (29.9%), and Microsoft (13.1%)(1999 figures), and each is vying with the others for domination of this crucial technology. In 1999, Microsoft grew at over 22% while Oracle and IBM remained fairly level. Thus Microsoft is growing at the expense of the smaller suppliers that are now down to less than 25% of the market.[9] Indeed, in a survey (Jan 2000) of enterprise data sources currently accessed by web servers, Zona Research found that 55 out of 81 corporations used Microsoft SQL server while 34 used Oracle, 24 used ODBC access to other formats, 24 used flat files and 15 used IBM DB2. This survey reflects the NUMBER of installations rather than the VOLUME of use. In general DB2 and Oracle tends to be deployed in high use applications while SQL server appears to be used in new and smaller applications. In 2003, IBM edged out Oracle as the market leader according to Gartner Dataquest:

Worldwide Market Share Held by RDBMS Software Vendors in 2002

Market Segment	IBM	Oracle	Microsoft
Total DBMS	37%	27%	18.8%
Total RDBMS	36.2%	33.9%	18.8%
Unix and Windows RDBMS	42.5%	24.2%	22.8%

Source: Gartner Dataquest, May 2003
http://www4.gartner.com/resources/115000/115036/115036.pdf

XML

Although XML is related to HTML, it is targeted at a different problem. That problem is enabling differing applications to recognize data embedded in documents even though that data may have different formats in the different applications. XML, which is short for eXtensible Markup Language, allows the use of extended tags that can tag content and identify it as specific kinds of data. The World Wide Web Consortium W3C, approved version 1.0 standards for the use of XML tags in 1999, and companies began to rush products into the markets. If the data is tagged properly and the applications know how to read the tags, then different kinds of applications can exchange data seamlessly. For example, a particular document may contain names and addresses of customers, but may also contain many other kinds of information. If the documents contain tags like <FirstName>, <FamilyName>, <City>, <State> or <HomePhone> then applications reading the documents can identify and extract the key pieces of information and use that information to populate a database or display data in a specific format.

Using XML enables this exchange, but it requires the development of a standardized set of tags that every one agrees upon. These standardized sets of tags are often referred to as “schemas,” and without them there is no guarantee that different applications will recognize one another’s tags. As so often happens when standards are required, there are now competing standards developing that will be largely incompatible with one another. Compatibility may be re-introduced through the development of “translators” or mappings between competing standards, but that is really putting a patch on a fundamentally flawed approach.

Microsoft Corporation is developing an ambitious set of XML standards under the BizTalk banner. BizTalk will likely contain over a 100 schemas for data exchange in various vertical market sectors. Another consortium of Sun, IBM Oracle and others (known as Oasis) is defining different schemas in overlapping vertical markets. There are also market specific consortia that are developing their own schemas for their own markets. One of the largest is Rosetta.net, a consortium of electronics suppliers and customers that are developing schemas for the electronics supply industry. There is also, ebXML, finXML, and a United Nations led consortium called UN CEFACT that promulgates standards for EDI and is translating those standards into an XML base.

EDI, Electronic Data Interchange, is an older enterprise approach to developing standards for exchanging data between unlike systems. XML is expected to extend the benefits of EDI to smaller enterprises and make it easier and more effective to do EDI in large organizations. Some observers expect that XML will essentially supplant or subsume EDI as a technology. "The EDI transactions were defined so concisely as to be unintelligible by humans," according to Fulton Wilcox, director of information management at BOC Gases, a supplier of industrial gases. "XML will allow us to exchange more human-readable data so that we can do a wider variety of things with our partners."

The problem with all of these standards is that there are so many of them and they may not be the same. If one group uses <FamilyName> while another uses <LastName> and a third uses <Surname>, then these applications will not be able to exchange data until a translator is developed that tells the applications that these tags all refer to the same thing. For XML to be as successful as it might be, there will have to be either industry consortia that can agree on standards for tagging data or there will have to be a few very powerful players who dictate standards to others in the industry, and who have the power to make those standards stick as defacto standards.

Soap

The Simple Object Access protocol (SOAP) is a way of using XML and http to contact objects on unlike systems and ask them to work together to solve a particular problem.

There are other protocols for interaction amongst unlike systems including COM, DCOM, and CORBA but SOAP is easier to implement across firewalls and may be better suited for lightweight applications. Soap encapsulates data and instructions into an XML described envelope using standardized encoding rules and RPC conventions, and then transmits that envelope for system to system through http. SOAP is supported by a collection of companies that numbers IBM and Microsoft among the leaders.

More on Java

As noted before, when the limitations of HTML began to chafe, Bill Joy, at Sun, adapted an idea from computing’s past. An early Pascal project from UCSD was designed to create a version of Pascal that was portable across platforms through the creation of an intermediate p-code that was then interpreted by platform specific interpreters. The goal was “write once and run anywhere.” Now Joy had originally envisioned this approach as a way of writing code for television “set top boxes,” an idea from the early 90’s that never really caught on. Searching for a use for his new baby, Joy suggested that this would be a great way to create an interactive “write once, run anywhere” language, Java, to be used either inside or outside of a browser. Using Java code inside a browser, would allow a server to deliver code on the fly across the network and run it on any platform. Because he knew that “real men do not program in Pascal,” he used a C like syntax over the Pascal like architecture.

Java also included many Internet specific capabilities and allowed easier use of object-oriented technologies, multi-threading of applications and automatic garbage collection of objects. Object oriented programming was a newer way of looking at programming as objects containing data and methods that responded to messages and events. Multi-threading allowed one to construct a program that would naturally undertake several tasks at once while keeping everything synchronized.

Garbage collection is a bit more difficult to explain. When a programmer is writing code, they need to create objects and data structures and these objects need to have a place in memory to live and work. Once the object or data is no longer needed, the object must be destroyed and the memory freed and returned to the system to be reused. In most languages, including C and C++, this must be done manually. The programmer creates the objects and allocates the memory for that object. The programmer is then responsible for determining when that object is no longer needed and then destroying the object and freeing the memory. For complicated programs this can be incredibly difficult to do. Mistakes in memory allocation and de-allocation are responsible for many of the “bugs” and crashes in commercial programs. If a programmer destroys an abject too soon, the program will crash when the object is needed and accessed. If the programmer fails to destroy the object and free the memory, then there will be a situation in which memory is allocated but never freed. Eventually this uses up all available memory and crashes the application and often the computer. This is often referred to as a memory leak.

Java included automatic garbage collection and excellent facilities for multithreading and Internet communication. Potentially, these technologies could improve both the development cycle time and reliability of the resulting code.

Sun advanced Java as a universal language that would essentially be put into the public domain and which would allow applications (Java Applets) to be delivered to the user “just in time through the network.

Java attracted strong support from the computer industry and especially from those companies that saw Java and the Internet as a way to change the game and wrest control from the industry domination enjoyed by Microsoft. If applications were to be truly “cross platform” and able to be delivered on the fly through the network, then it would not really matter whose operating system and whose applications they were. Microsoft had established industry dominance in both operating systems and office productivity applications and had used the close linkage between applications and operating system to both defend and extend that dominance into other areas. Java and the Internet might open a window of opportunity for others in the industry with other operating systems and other applications and might eventually free the computer industry from Microsoft’s dominance. It was not surprising then that IBM, Oracle, Netscape, and others invested heavily in Java, with over $100 million invested in a Java fund by 1996.[10] By the fall of 1997, IBM had more than 2,500 Java developers working around the clock around the globe.[11]

This vision was hampered by events both inside and outside the Java alliance. In a surprising move, Microsoft agreed to license Java and immediately set about incorporating Java into its products. Any feeling of triumph for the alliance must have been short lived, since it quickly became clear that Microsoft intended to “Windowize” Java by adding proprietary extensions and incorporating links to Windows only functionality. When Sun initiated legal proceedings against Microsoft, the dispute had the effect of chilling further investments by other organizations and developers who did not wish to invest time and money in a technology with a cloudy future. This was further complicated when Sun began to back away from a commitment to turn Java over to a standards body, choosing to retain control of Java development within Sun. Although Sun reiterated it’s commitment to an “open” Java, it’s alliance partners were not supportive of the move.[12]

Java was also encountering technical problems that slowed deployment. Since Java was designed to compile to an intermediate code and have that code interpreted at run time, it was destined to run more slowly than code written in C or C++. Also, using Java on a client machine meant that the Java Applets had to be downloaded when needed and this added further delay. To users accustomed to using fully complied ready-to-run applications, Java was insufferably slow.

There were candidate solutions for each of these problems, but they often added further complication. “Just-in-time” compliers could take the intermediate code and compile it immediately to improve the performance, but these came in several varieties with varying incompatibilities and once again Microsoft seemed to dominate the field. Others turned to Java on the servers as a way to write powerful portable applications quickly.

There were also alternatives to Java that slowed the development. JavaScript was easier to write and could do some of what Java promised to bring interactivity to browser-based applications. Legions of programmers who knew C and C++ coupled with the performance advantages of those languages has kept Java from large scale use for applications. Newer technologies such as Active Server Pages and XML have further eroded the need for Java.

In early 2001, Sun and Microsoft finally settled their lawsuit.[13] Microsoft agreed to make no new use of Java in any of its products and paid Sun $20 million. They retained the rights to distribute older versions without further modification. Microsoft proclaimed that it had no further interest in Java and that it would build its new .net applications in XML and using a new programming language called C#. Analysts differed on their perspectives of how much this might hurt the deployment of Java with many proclaiming that this was not an obstacle at all with others noting that it could be a major setback for ubiquitous use of Java.

Java has established itself as an important technology in the marketplace, but the overall vision for Java as a ubiquitous, open, portable, write once and run anywhere technology has not yet been fulfilled. It may never be. Java might indeed become just one of the many tools in the Internet developers arsenal to be deployed in those cases where it strengths give it an advantage. In effect Java has become a server side tool that has been adapted to create a framework for web services. Microsoft has countered this platform with it’s own set of tools called .NET.

JavaScript

JavaScript was designed a scripting language to be used to allow interactivity within a browser. It was far less powerful than Java, but it was also far easier to use and implement. JavaScript’s could be written with simple text editors and were simply included as part of the HTML code. There was no need to compile to intermediate code, then download, and interpret that code at run time. The use of the term “Java” within JavaScript is really more of a marketing than a technical description, since JavaScript was developed separately, is quite different, but was introduced during the peak period of Java hype. JavaScript continues to be heavily used to put interactivity into Websites, but it can introduce security problems and can be used by Web Sites to do various things on a user’s computer that can range from annoying to dangerous. It is often used by sites to “capture” browsers. Once you enter a site and then try to leave, it can keep redirecting you to other sites under the control of the original site. Often these are designed to keep you trapped in a never-ending loop of sites. At times the only way out is closing the browser. Sometimes that does not even work!

Because of the undesirable behaviors enabled by JavaScript, browsers allow you to turn scripting off, and many enterprises make it a practice to do so. This reduces the functionality of interactive sites, but improves security for the user.

ASP: Active Server Pages

Another way to personalize the web browsers experience is to provide a customized page for each browser. This requires that information be obtained from the user either through reading a cookie, tracking responses, or through asking for information. Then each page can be constructed especially for the user and delivered to that user alone. Active Server Page technology enables this practice and is a part of most commercial sites. Active server pages help allow the “mass customization” that can be an important part of the Internet experience. Each user sees a page that was designed for her and her alone. Java offers a competitive type of service on Web Servers other than Microsoft. The Java Server Pages (JSP) approach is often used on Unix and Linux based servers. These technologies, JSP and ASP, provide the core of the two competing web services architectures.

Cookies

One of the most controversial and debated technologies of eBusiness is the Cookie. The Cookie was invented by Netscape Communications to solve a problem that appeared to be a serious limitation of the World Wide Web and browsers that used HTTP, the Hypertext Transfer Protocol. HTTP was originally designed to be a “stateless” protocol.” That means that each transaction with a server or servers was independent. A server could not determine who you were, whether you had been there before, or what you had done previously. For eCommerce, this is a severe limitation. Web retailers really wanted to have a way of preserving information about transactions from particular machines. Netscape invented and patented the cookie as a way to store such information on the users own computer.

Perhaps Netscape intended that such information only reside there for the duration of a transaction. If so, then a consumer could fill a shopping cart with stored information in a cookie and then check out with all the information. The cookie could then be destroyed to ensure consumer privacy. Instead, the cookie stores information about the user that is not destroyed at the end of the transaction and becomes a unique identifier of the user on the particular computer. Although a cookie does not specifically tie actions to a named person, it does store a lot of information about whoever is using the computer.

The advantage of cookies for eBusiness is that each returning user is automatically identified to the system and thus a customized experience can be presented that takes advantage of the users prior activities on that machine and at that site.

Cookies are set by specific sites and on specific machines, but even that distinction began to break down when advertising suppliers, such as DoubleClick, began to provide banner advertising to many different sites. DoubleClick then controlled the ability to set cookies on the computers that could combine information about the consumer’s experience on many sites rather than just a specific site. As we shall see later, in the chapter on privacy and ethics, merging this information with databases tagged with actual consumer names and addresses could extend this further.

Cookies are often set by banner advertising. Several large web advertising companies, DoubleClick, Engage, and MatchLogic, set cookies for many web sites and can cross correlate action on all of the sites. Some consumer advocates have cited the use of very small graphics as a way of setting cookies without even alerting the consumer in the way a banner ad does.[14] These “transparent gifs” are really just single pixel graphics.

Consumers can easily turn of this capability by configuring their browser to not accept cookies, but the vast majority of Internet users have no idea how to do that. Even if they can, they then often find that it is difficult to complete web transactions without constantly reentering information. The privacy-convenience tradeoff often leads to the consumer giving up a bit of privacy for convenience.

Application Service Providers

Application service providers (ASP again!) provide application to users over the network. The concept is to rent an application rather than to buy it. Be careful here since ASP can refer either to Active Server Pages or Application service providers and they are very different things. We will cover the development of application service providers and web service more fully in a future chapter. It is only through the relentless advance of Moore’s Law and the Bandwidth law that we have arrived at a point where it might indeed be practical to download applications just in time as they are needed. This requires reliable network protocols, available bandwidth, and suitable technologies such as Java and XML to allow this to happen. Many think that the future of applications is renting rather than buying. One could look at this as the end of the old cycle of buy and upgrade, upgrade, upgrade or as the beginning of the era of continuous upgrades.

Connecting to the Net

Getting connected to the net offers many alternatives today. Among these are:

· POTS -Plain Old Telephone Service (14.4, 28.8, 56K ->)

· ISDN 64 or 128 kbs

· xDSL typically up to 1.5 Mbs

· Cable Modems (Roadrunner etc.) typically up to 1.5 Mbs

· Satellite (One way: Direct PC; Two way: Gilat to Home) and

· Wireless 802.11(b) is 11 Mbs and cellular is presently restricted to quite slow rates.

POTS remains the most likely connection from homes, but offers only low bandwidth dial up connections. Although 56 K modems are now widely used, it is quite rare to be able to establish a connection at rates of 40K or better. Those that want better bandwidth or “always on” service usually turn to one of the other options.

Wireless

"In the next two to three years, there are going to be more wireless ports than wired ports in the world. Cell phones are becoming an office with all the features of an office available around the world, and we are seeing the beginnings of this in the United States."

--George Nolen, president and CEO of Siemens' USA enterprise networks.

	Cell Phone Subscribers Year-end 1999 In millions
Vodafone Group	49.2
China Mobile	34.0
NTT DoCoMo	29.0
Verizon	19.8
France Telecom	19.7
T-Mobile International/ Deutsche Telekom	18.6
SBC	14.9
Telecom Italia Group	14.7
AT&T and AT&T Wireless	14.0
BT/BT Wireless	13.2

http://www.infoworld.com/articles/hn/xml/00/05/10/000510hnsiemens.xml

The future of eBusiness is undoubtedly wireless. Which protocol, which operating system, and which end user device are all open questions. Many think that one only has to visit Japan and hang out with the teenagers to discover the future of “always on” wireless connectivity. The DoCoMo system there has teenagers constantly chatting, paging, exchanging, and emailing in a system that is live as long as the battery is charged! DoCoMo, which stands for Do Communications Mobile, is an arm of NTT with over 10 million subscribers out of NTT total of 29 million. This dwarfs competitors in the United States, who have nothing similar to offer.

Wireless application protocol, hopes to create a formal standard that will allow wireless access across a variety of cell phones and manufacturers, but the Palm Pilot has already gone ahead and introduced wireless access with their own system.

The future is clearly cellular wireless, but beyond that it is difficult to say just how this will turn out.

Cellular wireless deployment is one area where the US is clearly not in the lead. As the accompanying table reveals, the leaders in cellular wireless development are in Europe and Asia.[15] Worse yet, the standards that have been most widely adopted in the United States are not those that the bulk of the world has chosen. Europe and many other areas have chosen an implementation of CDMA (Code Division Multiple Access) called GSM, that is not widely found in the United States.

The 802.11b wireless protocol (WiFi) has become something of a fad among users in the United States. It allows data transfer speeds of 11Mbs. It has become widespread as it is both cheap and reliable. Since WiFi requires access points connected to a wire network and only works over distance of the order of 300 feet, it is perfect for home wireless networks and small offices. WiFi is beloved by home users and reviled by enterprise IT administrators since it’s week security model and default wide open conditions can allow unauthorized user to access the network. None of this concern over security, much of which is overwrought, has dampened the enthusiasm of users. In fact one can download software such as NetStumbler, which will allow one to seek out and connect to networks that are left open for users. Perhaps the motivation is similar to open source advocates who believe that software should be free. Wireless advocates often feel that bandwidth should be free. WiFi is also made available by many hotels and even Starbucks, but they manage to get a toll collector on this information highway. Groups of wireless advocates have created organizations such as the “WARWalkers” who move through neighborhoods and map areas of wireless coverage. Some WARWalkers in cities have even begun to mark wireless access points in chalk on buildings and sidewalk. These “WARChalkers” allow the desperate WiFi user to find places that they can get connected. The author confesses to being addicted to wireless and enjoys the access that is now being provided in some many locations.

There are other protocols for wireless use such as 802.11a and 802.11g that provide even more bandwidth. The latter standard (g) is beginning to catch on in the United States since it operates in the same 2.4 GHz band and often provides compatibility with the (b) standard. The (a) standard has not attracted such a large following since it operates in different band and used quite different technologies. The actual technical details of the standards are as follows:

802.11 -- applies to wireless LANs and provides 1 or 2 Mbps transmission in the 2.4 GHz band using either frequency hopping spread spectrum (FHSS) or direct sequence spread spectrum (DSSS).
802.11a -- an extension to 802.11 that applies to wireless LANs and provides up to 54 Mbps in the 5GHz band. 802.11a uses an orthogonal frequency division multiplexing encoding scheme rather than FHSS or DSSS.
802.11b (also referred to as 802.11 High Rate or Wi-Fi) -- an extension to 802.11 that applies to wireless LANS and provides 11 Mbps transmission (with a fallback to 5.5, 2 and 1 Mbps) in the 2.4 GHz band. 802.11b uses only DSSS. 802.11b was a 1999 ratification to the original 802.11 standard, allowing wireless functionality comparable to Ethernet.
802.11g -- applies to wireless LANs and provides 20+ Mbps in the 2.4 GHz band.

(from Webopedia: http://www.webopedia.com/quick_ref/WLANStandards.asp )

The Bluetooth standard is yet another wireless protocol that is designed to allow devices to communicate on more of a peer to peer basis. Unfortunately it operates in the same band as the 802.11(b) and could lead to interference. It is a slower and shorter range standard that focuses more on communication among PDA or cell phone devices.