Throughout this document we will try to be consistent in our use of terminology. All of the various items available through the World Wide Web such as articles, pictures, files, etc. will be referred to as objects. The terms World Wide Web and Web will be used interchangeably, and they refer to the same thing. We distinguish between Web clients, Web servers, and push-cache servers. A server offers objects over the Web, and a client retrieves them. Servers push replicas of their objects onto push-cache servers. Push-cache servers are also servers, and may push objects onto other push-cache servers. Likewise, servers may also act as clients, retrieving objects from other servers. Finally, a host is any computer that is connected to the Internet: clients, servers, and push-cache servers included.
Throughout this document we will also make frequent references to the Internet; The Internet is a collection of many smaller networks, dubbed subnets. Subnets are connected to regional service providers that span specific geographic areas. Harvard, for example, is connected to NEARNet, the regional service provider for New England. These regional service providers are connected together by backbone networks: extremely high-speed networks that span the globe.