CONTENTS |
SUMMARY |
This report deals with the business challenge of competing with free software (also known as "open-source" software).
The two technologies in question, Perl and Apache, are described. A description of their competitive strengths is given. An analysis of their weaknesses and the opportunities that those weaknesses provide is given. A description of Sybase's current product strategy is given. Conclusions are drawn, and recommendations made.
It is concluded that the needs of enterprises will not continue to be met by Perl and Apache, and that Sybase must take steps to gain a dominant presence in the market segment in which Perl and Apache are currently the most popular technologies.
It is recommended that Sybase include the Perl syntax in PowerDynamo, offer a stripped down version of PowerDynamo for free to the public, and cultivate an online community for PowerDynamo. Simultaneously, Sybase must maintain its brand recognition as an enterprise-level solution provider.
INTRODUCTION |
With the stunning decision by Netscape Communications Corporation to release the source code for its Communicator suite earlier this year, and with the subsequent growing interest in free software, new challenges and opportunities are presenting themselves to the computing industry at large and, more imperatively, to companies developing software that may compete with such open-source products.
Sybase Incorporated is one such company, its PowerDynamo web application server shares, in some aspects, a market space with technologies such as the Perl programming language and the Apache web server.
This paper will analyze the current and future competitive environments of technologies in this market space. It will examine the benefits of free software and the opportunities they will provide in the future, and it will outline future strategies to take Sybase into a market-leading and profitable position.
This document will require familiarity with the concept of free software, the product PowerDynamo from Sybase as well as the following technologies: the Perl programming language and the Apache web server. Descriptions of these latter two technologies follow.
Perl is the most popular programming language currently used for web sites that need to execute operations on the server. Applications written in Perl are generally called via CGI (Common Gateway Interface) when a user makes a HTTP (Hypertext Transfer Protocol) request. Originally authoured by Larry Wall, Perl is currently developed and maintained by a volunteer community on the Internet. It is governed by The Artistic License, which allows anyone to use, distribute, and modify both the source code and the executable files with very few restrictions.
Apache is currently the most widely used web server in the world. According to the Netcraft Web Server Survey, which polled 2.2 million internet sites, Apache had a market share of 47.85%, more than doubling the next closest server, Microsoft IIS, which had a 21.27% share. Apache 0.6.2 was the first public release, released in April, 1995, developed by the Apache Group. The version publicly available at the time of this writing is 1.2.6, with 1.3 currently in beta. Version 1.3 will be the first version of Apache to run on the Windows OS (operating system). It's license allows anyone to use, distribute, and modify both the source code and the executable files with even fewer restrictions than The Artistic License.
ANALYSIS |
What makes Perl and Apache so popular? The easy answer, of course, is their initial purchase cost, which is zero (or the minimal cost associated with connecting to the Internet long enough to download them). But, just as it would be shortsighted to believe that the only people involved with free software are penniless hackers, so would it be shortsighted to believe that its price is the only reason for its popularity. While cost is a very important reason for the dominant market presence of Perl and Apache, there are a myriad of other, equally important reasons. As we look at these reasons, notice how each one fuels another, creating a network of benefits that attract and hold on to users.
To begin, free software is just that -- free. A user looking for a web server and for a language to make web pages dynamic can download Apache and Perl at no charge. That is all one has to do; there is no sales representative to call, no registration card to fill out, no deployment fees. What’s more, under the licenses of each product, distribution is unrestricted -- one friend can give it to another, it can be copied from the office computer to the home computer, one can even post it on one’s website for the whole world to download. This last scenario has happened hundreds of times, creating hundreds of mirror sites on the Internet and making the software both fast and easy to download. This unrestrictive distribution is one aspect of free software that breaks down the common roles associated with the marketing of commercial goods. It is not the traditional "producer to distributor to consumer" unidirectional model; with unrestricted distribution, every user has the opportunity to be a distributor. The paradigm shift does not end there: because the source code is open, the user can also take the role of a developer. The licenses allow users to make changes to the code to fit their specific needs, as well as to improve the code and redistribute it. These radical differences from how software is traditionally marketed causes the formation of an "online community" (postings in Perl newsgroups often refer to the "Perl Community"). Instead of a hierarchical structure, we now have something more like a kibbutz, where the entire community works together on a farm to produce the food that they all consume.
This community environment increases the value of the software by broadening the resources available to the user. Web pages and newsgroups abound with technical support, frequently asked questions, and the latest news -- all voluntarily contributed and managed by members of the community. Users are not limited to posing questions to a certain department of a single company, they have access to the entire community -- those who authour the software and know its inner workings as well as those using the software and have valuable experience. How do such resources compare to ones offered by commercial software companies? Very favourably, in fact, the Info World "Best Technical Support" award was bestowed upon the "Linux user community" in 1997 (Linux is a free software OS). In the case of Perl, members regularly offer and distribute applications and resources they have written in Perl to the community. For example, if a new user wants to add an online guestbook to his website, he doesn't have to code the whole thing because there are already Perl scripts written by other users for him to download.
This community environment is also the force which ensures the ongoing development (and hence, competitiveness) of the products. The developers of free software are generally not compensated financially, instead they are motivated by other factors, some practical and some psychological. The psychological factors are as diverse as the developers themselves. Obligation to 'give something back' to the community, status within the community, a sense of ownership, and straightforward altruism are among those psychological motivators. There are also compelling practical reasons for developers to participate. As mentioned before, the developers are also users of the software, so they have a vested interest in the continuing growth of the product -- if development were abandoned, they would have to endure the difficulties of migrating to new software. Another reason is free software's reliance upon and evangelism of industry standards. While commercial software companies attempt to create proprietary solutions that force a commitment upon the buyer to keep coming back to that one company, free software's open nature requires it to use open, industry-wide standards. This guarantee of industry standards is good news for the developers because it makes it easy for them to modify the software to solve the specific problems they themselves are facing, their audience will not be limited, and they do not have to wait for a company to release the next version to get the features they want.
One can see that there are many compelling reasons for people to take part in these software communities and that those involved are not just too cheap or too poor to pay for commercial software. There are, in fact, many competitve advantages for using free software like Perl and Apache.
Although Perl and Apache offer many advantages to users, they do not represent the ideal solution for everyone.
In the market today, Perl and Apache are positioned towards individuals and small- to medium-sized companies. Perl's syntax is more difficult to read than languages like Java and C++, making it hard for people other than the authour to understand and maintain -- a characteristic ill suited for large-scale and team-based development. Also, the language itself is optimized for scanning text files and extracting data, which may make it easier for individuals who store their data in text files, but it is far from ideal for organizations who store their data in databases. In the case of Apache, it is quite evident that the needs of large companies are not addressed. It is outpaced by commercial web servers in terms of how many HTTP requests it can handle per second (especially when a website needs to execute operations on the server side, as the only method Apache has for doing this is CGI), a very important factor for companies with large web sites potentially recieving millions of hits per day. It is also lacking when it comes to the latest security features, a great concern for companies doing any kind of commercial transactions over the web. In fact, several independant companies are making money by taking the Apache source code and adding the features that enterprises need.
This smaller-scaled positioning of Perl and Apache will create great opportunities for Sybase as the Internet's role in the activities of businesses grows. The world wide web is young, and the nature of the development being done thereupon will change as the needs of the enterprise begin to be addressed. Consider two key examples of where web development will be growing in the future: small companies who currently have web development as a focus, and enterprises who currently have small, fragmented web projects. As these small companies and small projects grow, they will require more complete and powerful technologies than Perl and Apache. When decision makers are ready to take that next step, they will be looking for technologies that will minimize their cost of migration and maintenance. Specifically, they will be looking for technologies that: allow them to reuse already written Perl code, web servers that deliver faster pages than Apache, tools that allow rapid development, and tools designed for team-based development. These are the needs of enterprise upon which Sybase must capitalize to acquire a market leading position in web technologies.
Sybase Incorporated has achieved brand recognition as a company that delivers open, enterprise-level solutions and products. It is important that Sybase maintain its positioning as enterprise-level, but to gain an industry-leading role in such a large market as web development, it must find a way to infiltrate the non-enterprise segment of the market as well.
As a product, PowerDynamo has two strategic objectives. One, to generate a revenue stream, and two, to generate sales of other Sybase products. The way in which it generates revenue is by offering advanced features such as database access and the ability to handle high-throughput websites that receive large numbers of hits per day. It is these two things that enterprises look for that Perl and Apache cannot deliver, and at $1995 US, it can potentially generate a good revenue stream by catering to this segment of the market. However, to truly capitalize on the switch that will occur as companies focus more and more on the web, Sybase needs to build greater product awareness of PowerDynamo and make it more accessible for current web developers. In effect, Sybase needs to construct a bridge between the world wide web of today and the web of the future.
CONCLUSIONS |
The popularity of Perl and Apache is due to: their low initial cost, their universal availability, the broad resources available to users through the software communities, and the guarantee of industry standards.
Enterprises looking for more robust technologies than Perl and Apache will need the following: an easier to read syntax than Perl, database access, fast, high-volume page serving, technologies that minimize the cost of migration, technologies that minimize the cost of maintenance, technologies that allow the reuse of Perl code, tools designed for rapid, team-based development.
To gain an industry leading role in web development, Sybase must infiltrate the non-enterprise segment of the market.
The key features of PowerDynamo which generate revenue are as database access and the ability to handle high-throughput websites that receive large numbers of hits per day.
To capitalize on companies who are increasing their focus on the web, Sybase needs to build greater product awareness of PowerDynamo and make it more accessible for current web developers.
RECOMMENDATIONS |
It is recommended that Sybase cultivate an online community for PowerDynamo. Steps that should be taken are:
In addition to DynaScript, the Perl syntax should be added to PowerDynamo, in order to allow legacy code written in Perl to be used by companies making the switch from Perl to PowerDynamo. The costs of adding this functionality to PowerDynamo could be minimized by leveraging the work already done on Sybase's product web.SQL, which used and interpreted the Perl syntax.
Sybase should continue to adhere to the EcmaScript standard with its DynaScript language and advertise that adherence as well as its commitment to open standards.
In order to infiltrate the lower-level segment of the market, broaden product awareness, and provide a bridge for companies to smoothly make the transition from free software to a Sybase web solution, Sybase should create a stripped-down edition of PowerDynamo and make it available for free or at a drastically reduced cost. Currently, Sybase derives revenue from PowerDynamo only because of the advanced features it offers, such as database access. If Sybase were to offer an edition of PowerDynamo without database access for free, it could gain a market-leading position with no losses in potential revenue. PowerDynamo's openness, ability to scale to low-end machines, and ease of use features (graphical user interface, templates) would make the stripped-down edition a strong competitor of Perl and Apache, as well as Microsoft's Active Server Pages (another technology available for no cost). As more enterprise-level applications need to be built, users of the stripped-down edition will find that their lowest cost course of action will be to upgrade to the fully functional edition.
With the release of a free, stripped-down edition of PowerDynamo, Sybase should make sure it maintains its positioning as an enterprise-level solution provider. It should emphasize in its marketing efforts that it is a large company acting as a patron to the entire free software community. Lessons about releasing free software while maintaining its enterprise-level recognition could be learned from Sybase's release of its free PowerJ Learning Edition product.
REFERENCES |
| The Artistic License | The Perl Organization | http://language.perl.com/misc/Artistic.html |
| Apache License | The Apache Group | http://www.apache.org/docs/LICENSE |
| 1997 Netcraft Web Server Survey | Netcraft | http://www.netcraft.co.uk/Survey/ |
| Best Of Awards | Info World | http://www.infoworld.com/cgi-bin/displayTC.pl?/97poy.supp.htm |
| Apache's Commercial Partners | The Apache Group | http://www.apache.org/info/support.cgi |
| The Cathedral and the Bazaar | Eric S. Raymond | http://www.apache.org/info/support.cgi |
| White Hats and Black | Tom Christiansen | http://www.apache.org/info/support.cgi |