Fabio Manganiello on Nostr: I still have concerns about the I/O overhead caused by having the full state of an OS ...
I still have concerns about the I/O overhead caused by having the full state of an OS regularly persisted to disk rather than memory. But the idea is intriguing.
#Databricks is working on the development of a database-based operating system (DBOS). Imagine having an OS where:
You can easily resume an application from a previous state, without having to restart it and replicate it.
The “it works on my machine”/“it doesn’t replicate” problems can be removed by design. If a process ended up in an unpredictable state, we have a relational db on the back to query exactly which was that state.
Backing up an OS and its current state becomes as easy as exporting a database dump.
https://developers.slashdot.org/story/24/03/18/0038232/database-based-operating-system-dbos-does-things-linux-cant
#Databricks is working on the development of a database-based operating system (DBOS). Imagine having an OS where:
You can easily resume an application from a previous state, without having to restart it and replicate it.
The “it works on my machine”/“it doesn’t replicate” problems can be removed by design. If a process ended up in an unpredictable state, we have a relational db on the back to query exactly which was that state.
Backing up an OS and its current state becomes as easy as exporting a database dump.
https://developers.slashdot.org/story/24/03/18/0038232/database-based-operating-system-dbos-does-things-linux-cant