TODO: write this document
As explained elsewhere, the information space is a virtual space. Each physical entity which "holds" information (be it a head or a sheet of paper), contains a projection of some of the information into it's data space. In other words Askemos is build on the assumtion that information is independant from the media, encoding and location of the data which contains it.
A projection of each place of the Askemos (blue) is kept in three servers (gray, red and yellow) of the network. These servers vote on all the messages, which arrive at the place (and have the capability to update the data of the place - so called write messages).
Emphasizing what it does, the software is a distributed operating system. Looking at how it fits together with other componentens it's an application server. If it was implemented at the "bare metal" - as usual for operating systems - we where left with the question who's implementing the applications. To escape the dilemma, a more practical approach was taken. The desired functionality is implemented as an application server, which can reuse all components available at the underlying host operating system.
Here structural comparison of an Askemos installation and a web server:
| user application interface | DSSSL, XSLT | CGI |
| Askemos | Apache / Zope | execution environment |
| network and | distributed file system | TCP / IP |
| persistent store | base operating system | network |
Information processing in that virtual space is done by a distributed virtual machine (DVM). That machine operates in the virtual space. Servers, which are interested in ongoing processing, support it: They keep a projection of the DVM, execute it's instructions to their best knowledge and vote among each other for the correct results.
| blue area | application level (information space, location independant) |
| read area | "own" physical server (user must trust it) |
| gray area | other server (not trusted) |