SlapOS aims for an efficient, decentralised, Free Software cloud with higher resilience and improved privacy. It is developed by a community including companies (Alixen, Bull, Dashing Soft, Nexedi, ViFiB, ZP Web Sites) and academia (Telecom ParisTech, University of Paris XIII)
The SlapOS model is based on standard Unix process/user separation. SlapOS can thus implement a wider range of Cloud models: virtual machine models, IaaS models, containers, instances, bare-metal PaaS among others. Nodes can be distributed geographically among different data centers and companies, or even into people's homes (self hosting). Software can also be deployed on more than 42 public clouds, even to create a Cloud federation where each infrastructure can back up another one. Billing is built into SlapOS through ERP5, the open source ERP, which acts as a process orchestrator and gives full independence. Other building blocks are Buildout to manage applications and supervisord for the monitoring. Documentation, forum and code repositories are available at community.slapos.org.
Mioga, the collaborative Extranet by Alixen, permits document sharing within inter- and intra-organisational workgroups through a fine-grained permission system. It complements an existing messaging system by providing groupware services: workgroup management, directory service integration, WebDAV or web-based document access, meta-data search, calendars, and many others. The source code (GPL) is available at www.alixen.org.
We introduce the main SlapOS concepts through simple examples using the command line: adding a node and requesting deployment of simple services like a database or a virtual machine in different regions of the world; then transforming our own computers, phones or some EC2 virtual machine into a SlapOS server and requesting deployment of ten LAMP applications inside of it.
We take Mioga as an example for how to port an existing, complex web application to the SlapOS cloud, be it for a SaaS offer or for internal usage (load-balancing and better use of spare servers). We can show code examples from all steps: from the Buildout "recipes" for components like PostgreSQL or Perl modules to the Slaprunner integration interface and recipes for both software and instance deployment. Administrative tasks such as supervision and backups can be fully automatised and integrated into the deployment recipes.
The SlapOS community would like to spread awareness and encourage participation. We conclude with an outlook of further work on security and robustness, welcoming questions and input of all kind.
Speakers: Cedric de Saint-Martin Viktor Horvath