Thursday, March 04, 2010

Chicago-01 Project

This project Idea is similar to
but with a more network aware setup.

I plan on having all managed code all in a virtual machine. The
virtual machine will boot up across a group of diskless / nondiskless
nodes. There will be no file system but more like central database
that all nodes can access and write to. I plan on having a nanokernel
on the diskless nodes that will load three programs
the virtual machine
synchronization of programs daemon
networking stack daemon
database client to send and receive data

The head node will have the management software
it will have the main database that all of the nodes read and write
to. I don't plan on having a filesystem of any kind

I was wondering what other ideas / comments you would have about this.

Posted via email from saltyschemer's posterous

1 comment:

Kohath said...

If you have no file system and you are going to 'load' more functionality (programs) from the database, how far are you going to align with that paradigm? Could you make the database functional higher up the list, and then load everything else from that, even some of the basic stuff like the synchronization daemon, etc.?