Discover. Understand. Anticipate.


Report written by Lorine Schieber

With RunAbove, gets ready for the FIFA World Cup 2014

Nicolas Capeyron created Sysadmin Badass back in 2009. His job: provide administration to web-oriented servers and build infrastructure architectures for his clients, e-commerce and tourism websites, or even SaaS providers. For, one of his clients, he chose a RunAbove public cloud. Here he explains why.

Nicolas Capeyron: I’m the web administrator of, a soccer website that has been hosted by OVH for many years. The site started to build up its audience and, with the UEFA Champion League final and then the FIFA World Cup in Brazil, our requirements exceeded the infrastructure’s capacity. Today we get 100,000 to 150,000 unique visitors per day, a number that can double or even triple during a popular event like the World Cup. Right now, we are getting 20 times more requests than the usual average as the Live SoFoot and the post-match analysis are very popular and well received. The ability to handle peak loads is thus critical to our business.

Did you turn to RunAbove to handle peak loads?

NC: At that point, for the World Cup or the European Championship, we would have migrated the site to a more powerful front-end server, then we would have the optimizations implemented. The operation was simple and not expensive; however, it had certain limitations – we could not ensure that the workload would stay within the server’s technical limits… The site didn’t have high availability or load balancing and, as we didn’t have much time to commit to development before the World Cup, we couldn’t have implemented but a few technical changes. Therefore, we needed a more flexible infrastructure in order to handle peak loads. I discovered RunAbove during the OVH World Tour in Lyon. I tried it, then bingo!

How did you integrate RunAbove to your architecture?

NC: We built this infrastructure to obtain availability and cost-effectiveness at the same time. The IP load balancing service dispatches the requests over five dedicated servers running HAproxy, which distributes the traffic towards another three different physical dedicated servers to which you can add five RunAbove VMs of different sizes. We thus combined the availability of IP load balancing, the OVH network, and HAproxy to link all our OVH and RunAbove front-ends. This allows us to handle the site’s 60 GB of static data very quickly.

Does RunAbove meet’s requirements?

NC: Yes, we are actually very pleased. The network and storage capacities are impressive. This is critical for Each front-end can now support the site, as well as a copy of the static data AND handle the additional traffic. In addition, by choosing OpenStack we can use classic DevOps tools such as Vagrant or Puppet. In short, we obtained excellent performance and an innovating cloud solution at an unparalleled price-performance ratio. If we had looked for a similar service elsewhere, we would have had to pay a lot more or rethink almost the entire site.