Basic Architecture

We'll look at the basic architecture for the web server part of Harry's game. An HTTP service as described above is used. It relies on a DNS server, provided by Harry, to maintain redundant sites, substituting new IP addresses for the site name as required. It also tries to use redundant IP address resources to minimise the need to do this.

Harry uses a web server to serve the content. The web server acts as the back end for the HTTP service, and is configured with war files which contain the content and harry's servlets. The Web server consists of a control system, which runs outside the HTTP service in its own Hosting contract, and back end netlets, which are created by the HTTP service and contain the actual web server.

Both web control system and web server back end netlets use a storage service. This also consists of a control system, running separately under its own hosting contract, and access point netlets, which provide the storage service to clients. The storage control system requires file space to store the actual data. This is obtained as a resource from the storage Hosting Contract.

The storage access points can use file system storage to cache data. This may be obtained through service connection between the Web Server and the storage system. The Web server may obtain the resource from the HTTP service through the Resource Discovery mechanism.

The Service and Resource relationships in Harry's game are depicted in figure .

Figure: Functional Overview of Harry's Game