This paper lays out the rationale for building a completely new operating system (OS) stack. Rather than build on a single node OS together with separate cluster schedulers, distributed filesystems, and network managers, we argue that a distributed transactional DBMS should be the basis for a scalable cluster OS. We show herein that such a database OS (DBOS) can do scheduling, file management, and inter-process communication with competitive performance to existing systems. In addition, significantly better analytics can be provided as well as a dramatic reduction in code complexity through implementing OS services as standard database queries, while implementing low-latency transactions and high availability only once.
I’m not even going to pretend to understand any of this.
It’s basically a system compromised of an OS and database with the fat trimmed off and optimised. It’s like a distributed form of PICK crossed with VMS crossed with DOS. Some specific features and abstractions are pervasive throughout the whole system which both simplifies code paths and enhances security and enhances parallelism. There may be other advantages too with a focus on SQL rather than complex programming languages. I suspect there are going to be disadvantages there too at a human resources level. Your recruitment pool will be similar in profile to a 1970s insurance company but then some organisations consider this a plus not a downside.
Let’s have a look at who sponsored the academic paper… Oh, why are none of them a surprise?
I wouldn’t be surprised if GCHQ had systems built like this. Mind you I wouldn’t be surprised if they didn’t either. A custom system running on custom hardware such has already been indicated in leaks can achieve astonishing levels of throughput compared to generic turnkey systems.
The entire system should use also less power and generate less heat per unit of useful output.
Reading academic papers and other primary sources is certainly less taxing on the brain than the tertiary fluff which passes for much of media and comment.
This sounds a lot like iSeries – “Database Distributed data management (DDM)”.
From the paper:
The source code, data, and/or other artifacts have been made available at https://github.com/DBOS-project
Remember WinFS?