An access control technique to limit access to information content such as available on the Internet. The technique is implemented within a network device such as a proxy server, router, switch, firewall, bridge or other network gateway. The access control process analyzes data in each request from the clients and determines if the request should be forwarded for processing by a server to which it is destined. Access control may be determined by comparing client source information against a database of Uniform Resource Locators (URLs), IP addresses, or other resource identification data specifying the data requested by the client. The invention therefore provides access control not based only upon content, but rather, based primarily upon the identity of the computers or users making the requests. The technique further avoids the problems of the prior art which categories or filters the content of only web pages based solely upon objectionable words. This is because a category database is used by the network device to control access and is created via a process involving human editors who assist in the creation and maintenance of the category database.