The 1996 Controversy about the 10-Connection Limit on TCP/IP in NT Workstation

by
January 1997

  • The Ten-Connection Limit in NT Workstation. July 1996. My open letter to O'Reilly customers urged them to spread the word that Microsoft's limits on the use of TCP/IP in NT Workstation was an attack on the freedom of individuals to publish on the web. Microsoft later responded by claiming that this policy was based on technical differences between NT Workstation and NT Server, a claim that was proven untrue by Andrew Schulman and Mark Russinovich. In followup, Microsoft backed down to an extent, by removing the technical limit, but keeping it only as a license restriction. Andrew Schulman provided further background on the controversy here. Unfortunately, many of the links in his article no longer work.

  • A copy of my Open Letter to Microsoft, and Microsoft's press release backing down a few days later. Internet.com's Intranet Design Magazine conveniently includes the text of the open letter I wrote to MS, followed by a copy of their July 19, 1996 press release announcing the removal of the technical (but not licensing) limits in question. Their editorial gives me credit for the change, which may be a little overstated, since a lot of people were ticked off about this.

  • In August 1996, I wrote a letter to the Justice Department, which was published online as well.

    In order to force people to use their product, Microsoft is now telling people that it is violating their license to use web technology (except in an extremely limited way) on any platform but the one on which their server is bundled.

    Not only is this bad for all Microsoft's competitors in the web server arena, it is extremely bad for the World Wide Web as a whole, since it institutionalizes a vision of the web that is extremely limited. In my email to Bill Gates and to Jim Allchin (copies attached), I pointed out that the Web is still in its infancy, and to set arbitrary limits on how users should deploy it, based not on advantages to users but purely on the basis of damaging competitors, could cripple this technology before it ever has a chance to reach its full potential.

    I would make the further point that Microsoft's claim that they are able to set licensing terms for the TCP/IP protocol (which, as an interoperable networking protocol, applies not just to their software but to any internet software) is a subtle and dangerous extension to their licensing rights. It transforms a licensing agreement covering a single company's products into a licensing agreement for an entire family of technologies--technologies, moreover, that Microsoft did not develop and has no proprietary interest in.

  • A brief summary of my comments on CNNfn, August 1996:

    "Speaking on CNNfn's Digital Jam, O'Reilly & Associates president Tim O'Reilly said he was questioned earlier this week by Justice officials, and that he told them that Microsoft is artificially trying to keep some competitors' software from functioning properly on its desktop Windows NT environment.

    [...]

    "O'Reilly said he's concerned by the fact that Microsoft -- which by its own admission is somewhat of a late-comer to the Internet software market -- is trying to take control of the global network. 'They're doing all they can to take control (of the Internet). In the process, I think they're damaging it pretty seriously.' ... Microsoft started creating a showdown when it began bundling its Web server with Windows NT, it's high-end operating system. Microsoft is reportedly telling companies they can't use competitors' software on the NT workstation platform and have sought to limit the use of standard Internet protocols with their software. ...

    "'They're saying 'you have to use our platform the way we want you to.' The vision they have is contrary to the way people want to use it. They're saying we'll tell you how to use the Internet and saying the way we want you to use it is the way that benefits our revenue the most.'"
    --CNNfn, 8.22.96