Dedicated Cloud

dedicated servers

dedicated servers


Load Balancing IP




database cluster

Ceph cluster


message queue

Wimi: an infrastructure with high availability combining Dedicated Cloud and dedicated servers through vRack

Wimi is a collaborative platform that was launched in 2012
The founders of the French startup Wimi: Antoine Duboscqet and Lionel Roux
Overview of a shared calendar in Wimi (at the top, shared tasks)
Overview of project documents within the Wimi interface.
Overview of "AirTime" feature (videoconferencing directly in the web browser with WebRTC technology).

The Wimi, the online collaborative work platform was launched in 2012. It's used by more than 25,000 businesses to date and it aims to make teamwork easier by bringing together all the services needed for effective collaboration - such as synchronisation and document sharing (Wimi Disk), task management, calendar sharing, chat, screen sharing and video and audio conferencing (Wimi Air Time) - within one interface. And all of this is combined with access rights management adapted for professional use. To host this platform, the French startup built a high availability infrastructure at OVH that combines Dedicated Cloud and dedicated servers, via the vRack.

Key information

  • 70 servers: 8 physical servers + 63 VMs running on OVH's Dedicated Cloud (4 L Hosts) and on a locally hosted OpenStack cluster (connected to the OVH infrastructure via a VPN)
  • 60,000 users worldwide
  • 1,000 https requests per second
  • 4,000 SQL requests per second
  • 10 GB database
  • Storage capacity 33 TB (files replicated 3 times)


  • High service availability: the application must be available 24/7/365
  • Data security: the goal is to have no data loss
  • Horizontal scalability: the infrastructure needs to be able to be scaled up transparently and elastically adapt to the number of connected users
  • Data hosted on French territory: a restriction imposed by certain businesses/organisations and a marketing tool in a sector dominated by big American companies
  • Performances: a cloud application has little benefit if it lacks fluidity compared to a local application (need for a cache system)
  • Secure connection at the service level for users (https) and at the infrastructure level for administrators (two-factor authentication access to VPN)

Solution deployed at OVH

The infrastructure is highly available for several reasons. Number one: Wimi chose a natively high availability (HA) solution - in this case Dedicated Cloud - in which every component is redundant, from hosts and the power supply to the network connection. Number two: all the services are separate, each one being powered by a cluster of independent servers to ensure the infrastructure's resilience. Within each cluster, operational servers in the cluster temporarily absorb the additional load, so there's no impact if one or multiple machines malfunction. This configuration also guarantees the infrastructure's scalability, as you only need to deploy additional machines at the congestion points to increase the infrastructure's capacity. Puppet (which will be replaced by SaltStack in future) can make this process easier by automating the deployment and configuration of additional VMs. To further improve the infrastructure's scalability, the infrastructure can be scaled up from a polling model which regularly checks for updates, to a push model which uses technologies like WebSocket + node.js. The latter enables you to push the content towards users rather than waiting for browser requests when users reload a page or for javascript to process AJAX requests.

1 – Load distribution with 2 Load Balancing IPs

2 - Front-end (Nginx + PHP-FPM)

3 - Back-end/Cache (PHP-FPM + Memcached)

4 – Databases (Galera cluster)

5 – File Storage (Ceph)

6 - Message queue (Rabbit MQ)

7 - Monitoring and performance analysis

8 - Backups

9 – The infrastructure's internal network


“When we began looking for a hosting provider to build our Wimi infrastructure in 2012, we didn't have much choice. Only the big companies offered virtualised infrastructure solutions combining flexibility and performance. However, these were American companies, and we wanted to reassure our users about the confidentiality of their data. We then discovered OVH and its Dedicated Cloud solution. The advantage of Dedicated Cloud, on top of the fact that OVH frees you from having to manage the cluster of physical machines, is that resources are dedicated and we can add hosts and datastores as and when we need them. The arrival of the vRack was the clincher, as it allowed us to combine the best of both worlds: the cloud for flexibility, and dedicated servers for raw power. This is particularly useful for hosting databases which receive lots of requests. Without this private network, we wouldn't have been able to deploy technologies such as MariaDB Galera Clusters or Ceph distributed storage. We could've connected the machines with OVH's public IPs, but that would have affected the infrastructure's overall performance and raised safety issues. Our ambition for 2015 is to develop Wimi internationally and, more specifically, in the USA. We’re considering extending our infrastructure into BHS, OVH's Canadian datacentre, and duplicating it there.”

Thomas Lemarchand (@t_lemarchand), Information Systems Manager at Wimi.

OVH advice

Wimi's infrastructure is interesting on several levels. It combines virtual and physical resources to guarantee high availability (HA) and scalability, makes the most of OVH's latest innovations (Wimi was one of the first users of Dedicated Cloud, vRack, Load Balancing IP and FS-Max servers), and relies on the most cutting-edge technologies, such as Ceph storage, MariaDB Galera Clusters and ELK monitoring.

One recommendation would be to set up a mechanism to monitor and identify critical thresholds for automatically adding additional resources. This way, when you pass a load threshold for a predefined amount of time, you would be able to start and stop scripts on new VMs via the OVH API. The VMs would be configured via Puppet or another configuration management system, and then automatically integrated in just a few minutes into the clusters that need reinforcements. Nevertheless, Wimi is a frequently used work tool and load peaks can be predicted. The infrastructure, which is nearly entirely automated, can therefore deploy an additional resource in 30 minutes - a very generous timeframe. With that in mind, the mechanism isn't absolutely essential.

Static files (JavaScript and CSS codes) from the web application may also be stored on the CDN to reduce the load on front-end servers while improving the webpage loading time, even if these files represent just a small proportion of all requests to the infrastructure.

Infrastructure components

1 Load Balancing IP

9.99 ex. VAT/month

Buy now

Dedicated Cloud

Enterprise Pack L 2 hosts AMD + Storage 2 x 300 GB

561.00 ex. VAT/month

Additional on-demand resources

Buy now


Service included with the Dedicated Cloud and the Infrastructure range servers

(Galera Cluster)

3 Infrastructure range servers
Model EG-32

104.99 ex. VAT/month

Buy now

Store files with Ceph

3 Storage range servers
FS-MAX (Pay-as-you-grow)

349.99 ex. VAT/month

Extra disks

Buy now


2 servers from the OVH Storage range
Model FS-72T

389.99 ex. VAT/month

Buy now