4 ms·

DEOS: The distributed exokernel operating system

44 points by friedrich12 9y ago

vermilingua 9y ago
The only info on the readme has been copy/pasted directly from wikipedia. Surely if this is a novel implementation of the idea, the author could put in a few words to explain how/why? EDIT: Actually, it appears the bootloader (boot/main.c) is a straight copy from https://github.com/wh5a/jos https://github.com/wh5a/jos, a teaching OS from an OS programming course at MIT. Is DEOS just someone's homework?
- friedrich12 9y ago
  I'm building off my MIT project the final project is implementing ideas from MIT exokernel paper sorry i didn't say that earlier.
  - friedrich12 9y ago
    I talked with Dawson Engler who was a part of the original exokernel project he said I should first start with the MIT OS from 6.828 and then start implementing ideas from the papers. I still have a lot to do though.
monocasa 9y ago
What about it is distributed?
jnwatson 9y ago
I’m pretty sure the name Deos was already used for another OS. IIRC one of the interesting properties of the original was that a single “process” could span multiple processor domains.
- chubot 9y ago
  I thought that this was DIOS, which is a different project from Cambridge. I read a paper on it a few years ago and I really liked it. I just checked the website and it seems like it's dormant? https://www.cl.cam.ac.uk/~ms705/research/dios/ https://www.cl.cam.ac.uk/~ms705/research/dios/ Since their heyday the 1980s, distributed operating systems---spanning multiple autonomous machines, yet appearing to the user as a single machine---have seen only moderate academic interest. This is a little surprising, since modern data centers might present an appealing environment for their deployment. In this position paper, we discuss what has changed since the community lost interest in them, and why, nonetheless, distributed OSes have yet to be considered for data centers. Finally, we argue that the distributed OS concept is worth a revisit, and outline the benefits to be had from it in the context of the modern data center.
  - jnwatson 9y ago
    No that's a different one. I guess the whole deus -> d*os is too hard to resist.
openasocket 9y ago
This is basically just a stub OS. For something with "distributed" in the name, there should probably be a network stack... or at least (since this is an exokernel) some sort of networking library.
- friedrich12 9y ago
  See my reply to vermilingua

44 points by friedrich12 9y ago

The only info on the readme has been copy/pasted directly from wikipedia. Surely if this is a novel implementation of the idea, the author could put in a few words to explain how/why? EDIT: Actually, it appears the bootloader (boot/main.c) is a straight copy from https://github.com/wh5a/jos https://github.com/wh5a/jos, a teaching OS from an OS programming course at MIT. Is DEOS just someone's homework?

friedrich12 9y ago
I'm building off my MIT project the final project is implementing ideas from MIT exokernel paper sorry i didn't say that earlier.
- friedrich12 9y ago
  I talked with Dawson Engler who was a part of the original exokernel project he said I should first start with the MIT OS from 6.828 and then start implementing ideas from the papers. I still have a lot to do though.

monocasa 9y ago

What about it is distributed?

jnwatson 9y ago

I’m pretty sure the name Deos was already used for another OS. IIRC one of the interesting properties of the original was that a single “process” could span multiple processor domains.

chubot 9y ago
I thought that this was DIOS, which is a different project from Cambridge. I read a paper on it a few years ago and I really liked it. I just checked the website and it seems like it's dormant? https://www.cl.cam.ac.uk/~ms705/research/dios/ https://www.cl.cam.ac.uk/~ms705/research/dios/ Since their heyday the 1980s, distributed operating systems---spanning multiple autonomous machines, yet appearing to the user as a single machine---have seen only moderate academic interest. This is a little surprising, since modern data centers might present an appealing environment for their deployment. In this position paper, we discuss what has changed since the community lost interest in them, and why, nonetheless, distributed OSes have yet to be considered for data centers. Finally, we argue that the distributed OS concept is worth a revisit, and outline the benefits to be had from it in the context of the modern data center.
- jnwatson 9y ago
  No that's a different one. I guess the whole deus -> d*os is too hard to resist.

openasocket 9y ago

This is basically just a stub OS. For something with "distributed" in the name, there should probably be a network stack... or at least (since this is an exokernel) some sort of networking library.

friedrich12 9y ago
See my reply to vermilingua