There is no cloud, only other people's computers! But our computers are second hand and run on green energy. So we're better, in an ish way.
This documentation is intended for cloud administrators and describes the inner working of the recycled cloud: it is not intended for end-users (as in it won't help them using the public services). Public service documentation can be found on the left sidebar.
Note: this documentation assume familiarity with UNIX environment. We gladly help (and might provide guidelines for newcomers in the future) but remember that you're responsible for your own working environment.
The goal of this project is to build a federation of small hosting providers with sustainability and ethics in mind. Using recycled hardware (with a few exceptions due to technical concerns) is a requirement.
From a technical point of view, the 'network' is composed of small independent 'locations'. We want to make the deployment and maintenance of these locations as easy as possible.
Further discussions on this page.
|Code||Location||: Status||Operated By||Description|
|LNTH||Linthal, CH||Production||e-Durable SA||Colocation at datacenterlight/ungleich|
|LSNE||Lausanne, CH||Testing||e-Durable SA||Testing ground + soon to be production|
|FVIL||Froideville, CH||Being deployed||e-Durable SA||Waiting on BGP peering|
Computers are complex beasts, there's a lot of things here! Feel free to reach out to us if you have questions.
Setting up the base infrastructure for a location is detailed on a separate page.
A typical location is composed of two edge routers, one core server and an arbirtrary number of workers netbooting from the core node.
It takes a lot of components to run a 'cloud' - here's our take!
All locations are equal, but some locations are more equal than others. Each member/location is free to choose what services they offer, its public and pricing. Some tooling is globally maintained for convenience purposes. A customer-facing management UI will be available in the future.
Note: there are plans to move common services out of VMs to LXD-based containers, as on-service per VM is easy but very inefficient.