The future Hypervisor
 

Hello,
Has anyone here heard the name DuVisor?
Please take a look at the following article and share your thoughts:
https://arxiv.org/abs/2201.09652

Thank you.

I checked the source code GIT repo. It requires a X86_64 CPU with a minimum of 4 cores. That means the efficiency and security it provides is not available where it is needed most (2 core mobile systems/laptops) but should work on the newer and less limited systems. It leverages existing virtualization tools Docker and QEMU.

I might like to play with it, but find it less than appealing for serious server work. For minor play on desktop systems and non-production applications it seems more intriguing.



Perhaps were I to have a use case where it fit well I would be more impressed, but for me and at this time it looks like an interesting toy.

Hello,
Thank you so much for your reply.
Is it a new technology?

It is built using RUST. Rust was born starting 2006, but the current form was released first in 2012. Many of the concepts and techniques were NOT new, but expressed in RUST a new way.

DuVisor expressed in rust in a similar way: using mostly old concepts and techniques to perform the function in RUST to operate in user space rather than kernel space. IT is not the first to use the concepts, or even the first to operate in user space, but at least ONE of the first to express the concept well in RUST.

Feel free to examine the code yourself, if you read RUST. IT seems free of corporate level suckage.

Hello,
Thanks again.
What is your opinion about the future of the Virtualization?

Virtualization is just too terribly handy to stop using, so it will have a place going forward as long as computing is an issue. I claim no crystal ball: I am the guy that said that no idiot would ever want a mechanical rodent on his desk, and here we are.

How MUCH place it will have, and if we will survive another century to make USE of it, are still at question.

The OP could also try a platform likeVirtuakBox or libvirt with virtmanager.

These days especially, there are many “good” solutions. But each one is definitely tailored for a different use-case. Many web-sites have already undertaken the task of canvassing the then-available options and holding them up against various usage scenarios. Don’t skip or shortchange this research.

Didn't read the paper, but I did look at the GitHub page:

https://github.com/IPADS-DuVisor/DuVisor

All this is, at this point, is a proof of concept.

Where did they, or anyone other than you, claim that this was the future of virtualization?