First someone does press a button.

Then his browser send a request to the BBC.

The BBC WEB font end server decides the request if for the DNA cluster

One of the DNA WEB servers catches the request.

All skins and sites get a BBC top bar header.

This DNA WEB server will collect the resulting page and will send parts of it as soon as 'ready'. Perhaps the page has been requested in the last minute and is still in cache?

Now comes the big magic part.

The central DNA server starts processing, it figures out the skin/site and decodes the request for the first selection.

The request can be a thread listing, an entry, a personal space, or an other selection.

First is checked on the DNA database for a redirection of skin/site and for Article number to User space.

Then the requested content for the page is collected at the DNA database.

Lots of subject and time stamp requests for listings, some code convertion for entries.

If all else fails an appropriate error message is requested from the database.

This htm document is replyed to the user.

The browser gets the page and does some requests for images.

< 10001101001011001010100100011110001101001011001010 11000110100101100101010001111000110101000111100011010 10001111000110100101100101010001111000110101000111100011010 11000110100101100101010001111000110101000111100011010 >

Somewhere between the browser and the DNA servers can be some kind of cache. Serving you the images if they are available within a (very) short time. (Well, this supports my experiences)

< 1010001111000110100101100101010001111000110101000 001101001011001010100100011110001101001011001010 10010001111000110100101100101010001111000110101000 01111000110100101100101010001111000110101000111100011010 >

The browser reveives the images and 'puts' them on the page (caching the non visible part of a page).