Apache Web Server

The Apache web server was born in 1995, but its history begins two years earlier with NCSA HTTPd. In 1993, Rob McCool developed the first portable HTTP server at the National Center for Supercomputing Applications at the University of Illinois. Distributed free of charge with its source code, the software enjoyed immediate success. Website administrators adopted it en masse, preferring this free solution to Netscape’s commercial servers, which cost several thousand dollars. NCSA HTTPd was then the most widely deployed server on the Internet.

But in 1994, the original team left NCSA to join Netscape, McCool included. Development came to an abrupt halt. Users found themselves with frozen software—functional, yes, but improvable. They had grown accustomed to fixing bugs themselves, adding features according to their needs, and then sharing these modifications on mailing lists in the form of patches. Brian Behlendorf was one of these active contributors; he notably created a password authentication system for the Hotwired website.

Faced with the project’s abandonment, a small group of webmasters decided to take up the torch. In February 1995, eight volunteer programmers including Brian Behlendorf and Roy Fielding created the “new-httpd” mailing list to coordinate their work. Two months later, in April, they released their first version: 0.6.2, built on NCSA HTTPd 1.3 and enriched with patches developed by the community.

The origin of the Apache name remains contested to this day. The foundation has long explained that it paid homage to the Apache Native American tribe, known for their endurance and warrior skills. Yet between 1996 and 2001, the project’s FAQ offered another explanation: “a patchy server”, a direct allusion to the numerous patches applied to the original code. Brian Behlendorf has successively validated both versions, suggesting that perhaps they both coexist.

Apache 1.0 arrived in December 1995. Robert Thau led this complete code rewrite, internally dubbed “Shambhala”. This version introduced a modular architecture that changed everything and simplified the addition of new features. Development became decentralized, innovation came from everywhere in the community, and results followed swiftly. By April 1996, Apache was the most widely used web server on the Internet. It has never relinquished that top position since.

The project drew inspiration from the IETF’s operating principles and its motto: “rough consensus and running code”. The rules remained simple. Anyone could contribute, but only core developers voted on official releases. Entry of new voting members required a nomination and unanimity among existing members. This system preserved the project’s coherence while remaining open.

Apache prevailed through its technical innovations. Virtual hosting, available from the summer of 1995 onward, addressed a pressing need for Internet service providers. This feature could manage multiple websites on a single server, up to 10,000 sites on one machine. Commercial solutions offered nothing comparable. The accessible source code provided another decisive advantage: users modified the software according to their needs, without waiting for a publisher to deign to implement a new feature or fix a bug.

IBM took interest in Apache in 1998. The computing giant was seeking a web server for its Websphere line. Rather than its own solutions, it chose Apache, recognizing its technical superiority. IBM understood where the value resided: not in the server, but in services and proprietary extensions like e-commerce systems. This collaboration marked the beginning of a lasting relationship. IBM in turn contributed to Apache’s development, enriching the project with its expertise.

The project’s structure took a new step in 1999 with the creation of the Apache Software Foundation, a non-profit organization. This evolution resolved legal and administrative aspects while preserving the project’s independence and collaborative culture. The ASF gradually developed a portfolio of projects, including Tomcat, the reference implementation of the Java J2EE architecture.

The core development team numbered about fifteen people, spread across the United States, Great Britain, Canada, Germany, and Italy. Contrary to stereotypes, these developers were not passionate teenagers but seasoned professionals: PhD students, doctors of computer science, experienced developers, company executives. Core developers produced about 80% of new features, but more than 400 people contributed to the code and over 3,000 reported issues.

Apache dominated the market despite fierce competition. Microsoft integrated its Internet Information Server into Windows NT as early as 1996, never exceeding 30% market share. Other players positioned themselves in niches, Zeus for very high-traffic sites for example. Some disappeared, like Netscape’s server, which had nevertheless reached 10% of the market.

In September 2009, Apache served more than 54% of global websites and more than 66% of the most visited sites. This dominance became self-sustaining: widespread adoption generated feedback and contributions that strengthened the software’s quality and stability, in turn attracting new users.

This success inspired countless other projects and shaped open source development practices.