From Computer Business Review, a sister publication

Stan Lepeak, a program director with market analyst the Meta Group, believes that there are two ways of looking at Netscape Navigator, the market leading Web browser. Either it’s the best piece of software that was ever written. Or it’s actually a pretty basic piece of code which is only good for displaying badly designed graphics pages, says Lepeak.

Looked at dispassionately, the view that Navigator is less than a remarkable software engineering achievement could be well founded. Navigator does not help the user send email, it cannot perform mathematical calculations, and it does not offer much in the way of text editing. In fact, it does not even do a particularly good job of helping users find their way around the Internet, something most people employ a specialist search engine to help them do. Yet, despite this simplicity and the product’s functional shortcomings, Navigator has established itself as one of the most successful computer applications ever. At the same time, it has helped to spark what threatens to be the next major revolution in the information technology (IT) business: the rise of the corporate intranet. In fact, between them, the success of World Wide Web-oriented programs such as Navigator, and the explosive emergence of intranet computing, are reshaping customer expectations of what IT can do, and what it should cost. In doing so, they are also causing the computer industry to seriously evaluate what it sells, and how it packages it. The emergence of the Internet and the Web-oriented computing paradigm has certainly been explosive. Earlier this summer Netscape announced that an estimated 38 million copies of Navigator are on desktops around the globe, a figure which it claims makes it the most popular desktop application in the world. The next most popular application, Microsoft’s Excel, can only claim to occupy 30 million desktops, and Microsoft Word occupies an even more modest 21 million.

Intranet servers

The growth in Navigator-equipped desktops is only matched by the growth in Internet servers. Server vendors such as Compaq now believe that Internet server sales may represent as much as 40% of the total server market in the US and Europe. What has only more recently been discovered, however, is that a great many of these ‘Internet’ servers are actually ‘intranet’ servers: communications platforms which use protocols and standards developed by the Internet community, but are dedicated to the internal company communications requirements. By 1998, predicts Zona Research, worldwide intranet server revenues will be worth $7.8 billion, almost four times the size of the $1.8 billion Internet server market. By the year 2001, says European market research group Input, intranet-related products and services will account for 20% of the entire IT product and services market. This represents a compound annual growth rate of 130%, double that predicted for Internet sales, and markedly higher than the growth in more conventional client/server systems technologies. These are spectacular projections, even by the standards of an industry accustomed to spurts of treble digit growth. They are all the more remarkable given the history of the Internet in the commercial sector which, until recently, was widely slated by vendors as too unreliable and insecure to be used in serious business systems. For a time, customers appeared to agree with vendors, and carried on with expensive investments in supposedly more sophisticated and robust client/server alternatives. Now though, user reticence is evaporating like morning dew in the sunshine. In the US, according to a poll of 50 major IT users by Forrester Research, 16% of major companies have already installed intranets, and a further 50% are set to follow suit, or are considering doing so this year.

Shining examples

Some of the pioneering intranet systems which have helped spark this rush of interest are spectacular successes, demonstrating the benefits of this technology which more complex re-engineering projects – that use expensive, state-of-the-art client/server technology – have failed to deliver. Federal Express’ Web-based tracking system is among the most famous of these. The global express delivery company took the unprecedentedly uncomplicated step of allowing its customers to trace the whereabouts of their parcels by tapping directly into a FedEx World Wide Web server. All Fedex’s customers needed to circumvent the inconvenience of dealing with a human intermediary between themselves and FedEx’s computer systems was a Web browser, a password and FedEx’s Web address. It proved a popular innovation, and is reputed to have boosted FedEx’s sales by as much $2 million.

Virtues of Intranet computing?

The FedEx Web system is not strictly an intranet. It pointedly does not exclude non-Federal Express employees from accessing the company’s Web-servers, but in most other respects it clearly demonstrates the virtues of intranet computing. Instead of requiring an expensive redesign and rewrite of FedEx’s conventional systems, the service exploits the company’s existing TCP/IP-based networks, and grafts a Web server onto existing databases. Through these simple expedients, confidential company information became simply and freely available to anyone with a Web-browser. The implications of FedEx’s achievement have not be lost on other IT users, or on their suppliers. Ever since client/server computing came into vogue at the beginning of this decade, it has been promoted as containing the keys which would finally unlock vast repositories of electronically-held data, putting that data into the hands of, or in Bill Gates’ phrase, at the fingertips, of those who can best exploit it. In practice, client/server computing has, at best, proved to be an expensive solution, and at worst, an ineffective one plagued by incompatibilities between the myriad of software components which need to be seamlessly connected for the systems to work. Now, says Meta’s Lepeak, the success of systems such as FedEx’s shows that we have a medium in which to do things we had hoped to do with client/server systems, or with Lotus Notes but couldn’t, either because of cost or complexity. Other analysts agree with Lepeak. Kathy Burrows at IDC has spent a year studying how companies in Europe have been approaching the Internet opportunity, and concludes that, while the uses that intranets are put to differ greatly from company to company, the underlying motivation for using an intranet remains consistent. People are doing it because it is simple and easy to implement, and requires minimum investment. You can do it on your existing servers, starting piece-meal and with minimal risk. It is a low-cost, low-risk investment which seems to be solving real problems.In the main, these real problems may not be so large or as potentially profitable as those tackled by FedEx. Indeed, the majority of intranet systems are devoted to ostensibly humdrum tasks such as providing online access to internal company telephone directories, publishing human resource manuals and forms, disseminating company organisational structures or posting situations vacant. These ‘brochureware’ applications may not reengineer the business exactly, but they do provide a cost-effective means of freeing essential information from the obscurity of company manuals. Once companies have established that publishing something on a Web site is a more effective way of keeping personnel up to date than swamping in-trays and trash cans with oceans of memos, they are only a short step away from expanding the breadth of intranets to carry information which is potentially more critical, and even to using it as a vehicle for executing transactions. It will not be long before users find out exactly how far the intranet/Internet paradigm can be stretched. With their ambitions fuelled by the market predictions of companies like Zona and Input, the IT industry is queuing up to apply Web-oriented technology to every conceivable application, and clearly vendors expect to line their pockets in the process. Oracle and Sun, for instance, are the first two major players to have clearly placed their bets on an Internet-oriented future. And, more recently, Robert Markham, executive vice president of Novell’s product group, predicted that intranet computer networks represent the next major stage in local area network computing. Intranets are the LAN/WANs of the future.

More than wishful thinking

There is more than an element of wishful thinking in the intranet-evangelism of these companies who stand to gain the most from anything which disturbs the Microsoft dominated status quo that currently exists on the desktop. However, they are not promoting intranet awareness in isolation. As well as a host of smaller companies, such as Verity and Open Text, which are pushing themselves forward as the vanguard of a new generation of intranet/Internet software companies, more established players are scrambling to establish an intranet/ Internet profile. Attachmate, the host connectivity specialist, now claims to be the market-leading supplier of client side applications for TCP/IP network environments, and even wants to be seen as a company which helped create the intranet phenomena. You can blame us for the Web hysteria, says Peter Zaballos, Attachmate’s vice president of marketing. Other leading client/server software players are less pretentious, but almost all are acknowledging the growth in intranet/ Internet computing. SAP plans to build Web links into its R/3 business software, and Lotus has similar plans for Notes, giving users a more simple way to interrogate and use the data held within their applications. Microsoft is building relationships with, and acquiring Web technology companies, such as Vermeer, supplier of Web publishing software, and now bundles a Web server capability with every copy of its Windows NT server operating system. Even systems integrators and service companies, such as EDS and Andersen Consulting, are now targeting their customers with made-to-measure intranet services, while BT and MCI have announced a joint strategy to provide globally managed intranet services for their customers.

Cash questions

Those without an intranet axe to grind, or a product marketing budget to spend, are a lot more circumspect about who, if anyone at all, will make a killing from intranet/ Internet related sales. Meta’s Lepeak points out that, so far, intranet computing has not been a product-led strategy. In many cases, it has proven to be quite the opposite – users have latched onto Internet technology as the raw material for a back-to-basics do-it-yourself approach to systems building. The magic formula has been that people have been able to roll out the infrastructure very quickly themselves, says Lepeak. Now EDS says, ‘you need an intranet, your IS can’t handle it, come and see us’. To build an intranet you don’t really need them. Good ideas are the best tools to use. The great thing about good ideas, from the customer perspective at least, is that they cost nothing. So, as long as users stick to doing nothing more ambitious than publishing brochureware over their intranets, many vendors will struggle to make significant capital from intranet projects. Indeed, it seems likely that before suppliers will begin to earn good profits from intranet-related products and services, intranets will have to play host to more sophisticated applications. Even then, suppliers may be forced to overhaul their pricing strategies if they want to compete successfully. At the moment, for instance, the relative expense of the intranet/Internet model when compared to a more orthodox client/server approach is often demonstrated by comparing the costs of a Web-based internal publishing system against that of a similar system based on Lotus Notes. It is an unequal contest. Typically, Notes systems cost between $1,000 and $1,500 per seat, per year for a medium-sized company implementation, with incremental client licenses costing around $75 each. Intranets are estimated to cost between $300 and $500 per seat, per year, with extra client software (a Web browser) costing as little as $25.

Relative simplicity

There are a variety of factors contributing to this price differential, most of which stem from the relative simplicity of Web-oriented products and tools, and of the systems which users normally build with them. A company which restricts itself to publishing a few HTML pages, for instance, places few demands on itself. Most IT staff with Unix experience will cope with HTML with minimal retraining, while users familiar with Windows will require virtually no tutoring in how to use a Web-browser. By comparison, to attempt to build a similarly basic system using Notes will require significant reskilling, both in the IS department and among the user community. The simplicity of first generation Web tools also impacts their cost in other ways. When browsers such as Netscape Navigator first appeared, they were functionally sparse, comprising only a few megabytes of code. This enabled companies such as Netscape to distribute their products very cheaply via the Internet, to the point where users were encouraged to download the software free of charge and only to license it once its use became significant enough to require the comfort blanket of being able to call on the supplier for support. This early ‘suck it and see’ distribution model enabled Netscape to establish a 30 million user installed base in under five years. It also set a trend for distribution and pricing which other Web-product vendors continue to adhere to. Today, purpose built Web products remain characterized by their rarefied functionality, and a pricing model which takes no account of the size of the installation the software will serve. This is even true of Web software supplied by established vendors. Microsoft, for instance, bundles its Web server software with Windows NT at no extra cost. The cost model for systems such as Notes is quite different, and with good reason. Unlike the bare bones functionality which typifies Web applications, client/server application platforms such as Notes come complete with a purpose-built distributed flat file database built in, and highly configurable security options. These, and a whole raft of other features make Notes a desirable development platform for sophisticated, industrial strength groupware and collaborative computing applications. When a company invests in Notes, it is buying into an architecture on which to found an ongoing program of distributed, collabarative computing systems development, in much the same way as, 15 years ago, a company which invested in an IBM mainframe system believed they were buying a development future for their core business systems. As long as this perception persists, customers may continue to pay premium prices for complex, proprietary application environments such as Notes, or SAP’s R/3, and stick to applying Web technology to non-strategic point solutions.

Strategic importance

This view of client/server systems somehow being of strategic importance, while Web-based systems are suitable only for tactical solutions, however, is unlikely to go unchallenged for long, particularly as the momentum behind intranet computing builds, encouraging more vendors to build links between their products and those that are appearing on the Web. Both SAP and Lotus, for instance, are taking steps to make information held in Notes and R/3 accessible to clients who may only be equipped with a Web-browser. Most of the database vendors are following suit, and are even importing Web-based search tools, such as Verity’s Topic, into their core relational databases. On a more general basis, the transfer of information from HTML-based forms into SQL-based relational databases (an essential step to speed the acceptance of Web-based forms as the starting point for transaction services) will become much simpler once the development of the Common Gateway Interface (CGI) standard is completed – which will probably be next year. These developments will make it easier to move information between Web-based clients and client/server servers, opening the installed base of existing transactions to a Web-oriented client base. The real threat to conventional client/server systems will emerge as more core functionality is implemented directly on top of the Web infrastructure, both through the introduction of server systems developed specifically for the Web, and with the emergence of Java-based applets which will allow Web clients to access and download business logic as easily as they now download Web pages.

Fast growing market

Although still small, the market for made-for-the-Web server applications is already growing fast. Some of the emerging products, such as Veriphone’s merchant transaction software, are highly specialized and address specific problems associated with utilizing Web software for transaction systems. However, other companies are entering the field with products which match the functionality of established client/server products, while still adhering to the price and distribution models which have been established by pioneer products such as Netscape Navigator. One such product is the LiveLink Intranet collaborative computing suite from Open Text of Canada. LiveLink is a compendium of applications offering a broad range of functions including search and retrieval, document management, project management and workflow. It is available in two flavors – a client/server version and a version which is specifically written for implementation across an intranet infrastructure. Customers will notice that price differentiates the client/server version from the Web version of LiveLink. The latter is available as a Windows NT application with individual functional modules priced at $12,500 for the first application, with further modules costing $6,250. At these prices, a complete LiveLink suite would cost around $30,000 with no ceiling on the number of clients that can be attached. The client/server version, written for Unix RISC systems is basically double that price says David Weinberger, vice president of strategic marketing at Open Text. However, Weinberger believes that there are more important advantages to the Web-version of LiveLink than simply its price. Compared with client/server collaborative computing products, he points out, LiveLink represents a a functional view, rather than an architectural one. This is a direct contrast to the approach taken by Lotus, for instance. It allows Open Text’s customers to buy the functionality they require, as they need it and without bearing the premium price of a highly architectured product such as Notes. New functionality can be added module by module, and because there is no requirement to adhere to a proprietary architecture, those modules need not originate from Open Text. Any application designed for implementation across a Web system can be plugged in as required, and Open Text already supplies components from other vendors as part of its Internet Anywhere option which includes Netscape Navigator, the Eudora Lite Web-based electronic mail product and SoftQuad’s HoTMetaL Web-authoring tool. The largest part of Open Text’s installed base is still accounted for by the conventional client/server versions of its products, but this is beginning to change, says Weinberger. Major customers, including Ford, started off piloting the software on their intranets but now, he says, they are looking at the second phase of these projects and are preparing to roll out the applications to populations as large as 5,000 users. In the future, he believes, the Web-based version of LiveLink will be the most important part of Open Text’s business.

Intranet impact

As Web-based applications grow to account for a larger, possibly dominant proportion of the total applications deployed by corporations, it will not only be suppliers of conventional business suites which will suffer. Suppliers of systems software could also find their positions undermined. A recent Forrester Research report, ‘The Full Service Intranet’ predicts that as more server services are delivered over intranets, these, as opposed to conventional network operating systems, will become the chief points of definition for the IT environment. Forrester’s view rests on the idea that core services, such as directory, electronic mail, file, print and network management will become standards-based features of tomorrow’s intranets, usurping many of the roles currently associated with products such as Windows NT and Novell NetWare. As with any other revolutionary proposal, at first sight, Forrester’s theory appears to be preposterous. How, after all, can something as amorphous as the World Wide Web spawn anything so robust that it can replace established market leaders such as Windows NT and NetWare?

The future is functional

Forrester’s view backs up Weinberger’s assertion that the future belongs to those who take a functional, rather than an architectural view, and is supported by other independent voices who point out that the Web browser – possibly the least sophisticated element of Internet technology – has already gone a long way to supplanting conventional desktop GUIs as the principal systems interface of many companies. The plug-and-play architecture of the Internet, which is alien to most GUIs, is fundamental to most Web browsers, says James Eisbisch, an analyst with Input. It is a characteristic which is set to be reinforced once the first generation of Java and Micro- soft Active-X applets arrive later this year, and may extend to physical device independence if the network computer initiative launched by Sun and Oracle is successful. With these prospects in store, it may not be long, says Eisbisch, before users recognise that the intranet paradigm may in fact contain the ultimate expression of the truly open systems market that has proved so elusive so far. It is far more difficult to get locked into an Intranet than it is to an operating system, he says. Ultimately, Input predicts that intranets will be delivered as value-added services provided by network companies, with business applications delivered in the same way that today’s value added network services provide guaranteed message delivery or network management. The timescale for this evolution of the intranet from an internal company publishing medium, to a complete business application delivery environment (four years in the case of Forrester, five years according to Eisbisch) is long enough for there to be plenty of unforeseen developments which may cause the pundits to revise their predictions. Irrespective of the basic nature of early Web software such as Navigator, it has helped spawn a computing paradigm which shows no signs of being a flash in the pan. As Open Text’s Weinberger argues the Web paradigm is proving persuasive, and that’s why it’s going to be pervasive.

Intranet building blocks

Intranets are made up of five key pieces of technology – browsers (the client interface); servers (Web server software); platforms (server and client operating systems); authoring tools (to build the Web pages); and firewalls (to keep the site secure from outside penetration). In browsers, the overwhelming market leader is Netscape with its Navigator product. According to market analysts Zona Research, Navigator has an 85% market share while the closest market rival, Microsoft with Internet Explorer, boasts just 4%. However, Microsoft’s decision to bundle Internet Explorer with its Windows 95 operating system, should see it share of the market grow. Netscape is also the clear leader in Web server software. According to Zona, it owns 71% of the market. No other single vendor makes a significant showing. In terms of server platforms, intranets, like the Internet, used to be dominated by Unix but Microsoft Windows NT has now caught up. According to Zona, the market is currently split nearly 50/50 between Unix and NT, although it expects NT to dominate by year-end. On the intranet client side, Microsoft accounts for nearly three quarters of the market with Windows 95 and Windows 3.1.The market for Web authoring tools is more evenly spread. There are four key companies, Quarterdeck, SoftQuad, Adobe and InContext, and each has roughly a 20% market share. Intranet firewalls is one of the hottest market segments of all, with a large number of vendors fighting for market share. Key players including IBM, Checkpoint Software, Network Systems Corp, Digital Equipment and SOS Corp.

By Phil Jones