Re:Replacing Linux with a Database System
by acqq on Sunday July 29, 2018 @04:10AM
(#17637065)
I read the paper completely differently than you. I think you were influenced by the title of the news article. The paper seems to me very reasonable:
Given the goal of implementing an architecture for part of the tasks of the OS for a 32,000+ core supercomputer, can we use mathematics to model such an architecture? The answer is yes, the mathematical model is defined and the simulation of that compared with the equivalent implementation provided in the current Linux.
The way I see it, it was then never about "replacing the whole Linux" but about "how to better implement some piece an OS which is to manage e.g. 32,000+ cores and simultaneous forkings on them."
Given the goal of implementing an architecture for part of the tasks of the OS for a 32,000+ core supercomputer, can we use mathematics to model such an architecture? The answer is yes, the mathematical model is defined and the simulation of that compared with the equivalent implementation provided in the current Linux.
The way I see it, it was then never about "replacing the whole Linux" but about "how to better implement some piece an OS which is to manage e.g. 32,000+ cores and simultaneous forkings on them."