forked from toolshed/docs.coopcloud.tech
159 lines
5.4 KiB
Markdown
159 lines
5.4 KiB
Markdown
---
|
|
title: Comparisons
|
|
---
|
|
|
|
We think it's important to understand that *Co-op Cloud* is more than just
|
|
software and technical configurations. It is also a novel organization of *how*
|
|
to [create technology socially](https://docs.coopcloud.tech/federation).
|
|
However, strictly technically speaking you may be wondering:
|
|
|
|
### What about `$alternative`?
|
|
|
|
We have various technical critiques of other similar projects which are already up-and-running in the ecosystem, as they don't necessarily meet our needs as a small tech co-op. However, Co-op Cloud isn't meant to be a replacement for these other projects.
|
|
|
|
Here is a short overview of the pros/cons we see, in relation to our goals and needs.
|
|
|
|
### Cloudron
|
|
|
|
#### Pros
|
|
|
|
- 👍 Decent web interface for app, domain & user management.
|
|
- 👍 Large library of apps.
|
|
- 👍 Built-in SSO using LDAP, which is compatible with more apps and often has a better user interface than OAuth.
|
|
- 👍 Apps are actively maintained by the Cloudron team.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Moving away from open source. The core is now proprietary software.
|
|
- 👎 Libre tier has a single app limit.
|
|
- 👎 Based on Docker images, not stacks, so multi-process apps (e.g. parsoid visual editor for Mediawiki) are a non-starter.
|
|
- 👎 Difficult to extend apps.
|
|
- 👎 Only supported on Ubuntu LTS.
|
|
- 👎 Upstream libre software communities aren't involved in packaging.
|
|
- 👎 Limited to vertical scaling.
|
|
- 👎 Tension between needs of hosting provider and non-technical user.
|
|
- 👎 LDAP introduces security problems - one vulnerable app can expose a user's password for all apps.
|
|
- 👎 Bit of a [black box](https://en.wikipedia.org/wiki/Black_box).
|
|
|
|
### YunoHost
|
|
|
|
#### Pros
|
|
|
|
- 👍 Lovely web interface for app, domain & user management.
|
|
- 👍 Bigger library of apps.
|
|
- 👍 Awesome backup / deploy / restore continuous integration testing.
|
|
- 👍 Supports hosting apps in subdirectories as well as subdomains.
|
|
- 👍 Doesn't require a public-facing IP.
|
|
- 👍 Supports system-wide mutualisation of resources for apps (e.g. sharing databases by default)
|
|
|
|
#### Cons
|
|
|
|
- 👎 Upstream libre software communities aren't involved in packaging.
|
|
- 👎 Uninstalling apps leaves growing cruft.
|
|
- 👎 Limited to vertical scaling.
|
|
- 👎 Not intended for use by hosting providers.
|
|
|
|
### Caprover
|
|
|
|
#### Pros
|
|
|
|
- 👍 Bigger library of apps.
|
|
- 👍 Easy set-up using a DigitalOcean one-click app.
|
|
- 👍 Works without a domain name or a public IP, in non-HTTPS mode (good for homeservers).
|
|
- 👍 Deploy any app with a `docker-compose.yml` file as a "One Click App" via the web interface.
|
|
- 👍 Multi-node (multi-server) set-up works by default.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Single-file app definition format, difficult to tweak using entrypoint scripts.
|
|
- 👎 Nginx instead of Traefik for load-balancing.
|
|
- 👎 Command-line client requires NodeJS / `npm`.
|
|
- 👎 [Requires 512MB RAM for a single app](https://github.com/caprover/caprover/issues/28).
|
|
- 👎 [Backup/restore is "experimental"](https://caprover.com/docs/backup-and-restore.html), and doesn't currently help with backing up Docker volumes.
|
|
- 👎 Exposes its bespoke management interface to the internet via HTTPS by default.
|
|
|
|
### Ansible
|
|
|
|
#### Pros
|
|
|
|
- 👍 Includes server creation and bootstrapping.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Upstream libre software communities aren't publishing Ansible roles.
|
|
- 👎 Lots of manual work involved in things like app isolation, backups, updates.
|
|
|
|
### Kubernetes
|
|
|
|
#### Pros
|
|
|
|
- 👍 Helm charts are available for some key apps already.
|
|
- 👍 Scale all the things.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Too big -- requires 3rd party tools to run a single-node instance.
|
|
- 👎 Not suitable for a small to mid size hosting provider.
|
|
|
|
### Docker-compose
|
|
|
|
#### Pros
|
|
|
|
- 👍 Quick to set up and familiar for many developers.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Manual work required for process monitoring.
|
|
- 👎 Secret storage not available yet.
|
|
- 👎 Swarm is the new best practice.
|
|
|
|
### Doing it Manually (Old School)
|
|
|
|
#### Pros
|
|
|
|
- 👍 Simple - just follow upstream instructions to install and update.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Loads of manual work required for app isolation and backups.
|
|
- 👎 Array of sysadmin skills required to install and maintain apps.
|
|
- 👎 Hard to share configurations into the commons.
|
|
- 👎 No idea who has done what change when.
|
|
|
|
|
|
### Stackspin
|
|
|
|
#### Pros
|
|
|
|
- 👍 Easy instructions to install & upgrade multiple tightly integrated apps.
|
|
- 👍 Offers a unified SSO user experience.
|
|
- 👍 Offers tightly integrated logging, monitoring, and maintenance.
|
|
- 👍 Has a strong focus and attention to security.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Upstream libre software communities aren't involved in packaging.
|
|
- 👎 It is not designed to be a general specification.
|
|
- 👎 Hard to share configurations into the commons.
|
|
- 👎 Significantly limited library of eight apps.
|
|
- 👎 Additional apps are treated as "External Apps" with only OAuth2/OpenID integration.
|
|
- 👎 Requires a Kubernetes cluster.
|
|
|
|
|
|
### Maadix
|
|
|
|
#### Pros
|
|
|
|
- 👍 Nice looking web interface for app, domain & user management.
|
|
- 👍 Offers a paid hosting service to get up and running easily.
|
|
|
|
#### Cons
|
|
|
|
- 👎 Upstream libre software communities aren't involved in packaging.
|
|
- 👎 It is not designed to be a general specification.
|
|
- 👎 Hard to share configurations into the commons.
|
|
- 👎 Limited library of apps.
|
|
- 👎 Uses *OpenNebula*, *Ansible*, and *Puppet* as underlying technologies.
|
|
- 👎 Appears to be only a team of two people.
|
|
- 👎 Appears to be inactive on Mastodon and limited GitLab activity.
|