4 ms·

Cuboid: An application-centric, decentralised and distributed computing project

41 points by tasos_laskos 3y ago

tasos_laskos 3y ago
I've created a project that's been published for a while now and now starting to think about making it known to folks. As the title reads, its name is Cuboid and is part of the Qadron org -- Ecsypno's < http://ecsypno.com http://ecsypno.com > F/OSS branch. In essence, the Cuboid Framework offers the possibility of easily creating decentralized and distributed applications in Ruby. In hipper terms, you can very easily setup your own specialized Cloud or Cloud within a Cloud. It's like a Ruby-on-Rails deal but for distributed and multi-faceted computing. Its features have up until now been designed to cover my company's products' needs but I'd love to get everyone's take on it and build a community around it to drive it further. The project goal is to make distributed computing fun and easy, it should feel like a usable toy. Cheers!
thunderbong 3y ago
How's this different from DRb[0] ? [0]: https://github.com/ruby/drb https://github.com/ruby/drb
- tasos_laskos 3y ago
  DRb is comparable to Toq[0], Cuboid isn't just RPC, it's an application framework. If you checkout the README[1] you'll get all the details. [0] https://github.com/qadron/toq https://github.com/qadron/toq [1] https://github.com/qadron/cuboid/blob/master/README.md https://github.com/qadron/cuboid/blob/master/README.md
rgbrgb 3y ago
Looks cool. I’m a big rails fan. Could you give an example of what kind of apps this might be useful for?
- tasos_laskos 3y ago
  Applications that could benefit from clustering would greatly benefit from this. This means apps that solve problems in parallel and that list is endless. What's cool is that you can distribute an app whose payload is in an extension (C/C++/Rust) and then use Cuboid for parallelism that spans machines quite easily -- think simulations. So the performance intensive part is handled by something high performance in its domain and then Cuboid/Ruby handle the clustering.
thdespou 3y ago
Why ruby and not Rust or Java BTW?
- tasos_laskos 3y ago
  It was incubated within a Ruby project I was working on. And extensions do exist so you can still get the same performance, but have clustering that is as accessible as Ruby.

41 points by tasos_laskos 3y ago

I've created a project that's been published for a while now and now starting to think about making it known to folks. As the title reads, its name is Cuboid and is part of the Qadron org -- Ecsypno's < http://ecsypno.com http://ecsypno.com > F/OSS branch. In essence, the Cuboid Framework offers the possibility of easily creating decentralized and distributed applications in Ruby. In hipper terms, you can very easily setup your own specialized Cloud or Cloud within a Cloud. It's like a Ruby-on-Rails deal but for distributed and multi-faceted computing. Its features have up until now been designed to cover my company's products' needs but I'd love to get everyone's take on it and build a community around it to drive it further. The project goal is to make distributed computing fun and easy, it should feel like a usable toy. Cheers!

thunderbong 3y ago

How's this different from DRb[0] ? [0]: https://github.com/ruby/drb https://github.com/ruby/drb

tasos_laskos 3y ago
DRb is comparable to Toq[0], Cuboid isn't just RPC, it's an application framework. If you checkout the README[1] you'll get all the details. [0] https://github.com/qadron/toq https://github.com/qadron/toq [1] https://github.com/qadron/cuboid/blob/master/README.md https://github.com/qadron/cuboid/blob/master/README.md

rgbrgb 3y ago

Looks cool. I’m a big rails fan. Could you give an example of what kind of apps this might be useful for?

tasos_laskos 3y ago
Applications that could benefit from clustering would greatly benefit from this. This means apps that solve problems in parallel and that list is endless. What's cool is that you can distribute an app whose payload is in an extension (C/C++/Rust) and then use Cuboid for parallelism that spans machines quite easily -- think simulations. So the performance intensive part is handled by something high performance in its domain and then Cuboid/Ruby handle the clustering.

thdespou 3y ago

Why ruby and not Rust or Java BTW?

tasos_laskos 3y ago
It was incubated within a Ruby project I was working on. And extensions do exist so you can still get the same performance, but have clustering that is as accessible as Ruby.