High availability

Disaster Recovery Planning

fault tolerance

ERP hosting

mission critical application

Multi datacenter infrastrucure

3 tier architecture

MySQL replication

Load Balancing IP

dedicated server

Backup server

So you Start


HA and PRA Infrastructure for hosting critical applications (ERP)

A computer engineering service company's customers include a firm with 70 expert accountants. To host the firms ERP, the initial choice was a Cloud VPS solution, because its SLA guarantees maximum availability of the application. Facing an increase in the number of ERP users, the Cloud VPS’s capabilities revealed to be insufficient. Migration to a dedicated server was required, but much thought had to be put into how to deal with any equipment failure. Ultimately, in the end, a high-availability infrastructure was deployed, based on 2 identical servers, in addition to a backup server that was added to secure data as part of a disaster recovery plan (DRP).

This case is inspired by the infrastructure designed by EURL Barme for its client, Sextant Expertise was the OVH 2014 Infrastructure Trophy winner in the category “Enterprise Infrastructure (outsourcing)”.

Key information

  • Hosting the ERP of a firm of expert accountants(mission critical application)
  • 70 daily users
  • 2 identical dedicated servers, one hosted in SBG and the other in RBX
  • Backup on a So you Start server
  • VPS Cloud for HA monitoring


  • ERP accessible from the Internet secured by an SSL certificate
  • High availability of the enterprise’s mission critical application
  • Data security: Data is sensitive (including financial information) and there can be no risk of data leak or loss

Solution deployed at OVH

When we have to host a critical application, the first reaction is to rely on two servers: one being an active server (master) and the other a passive server (slave), capable of taking over in the event that the master should fail. This solution is perfectly suited as part of a disaster recovery plan (DRP) and built around two metrics, RTO (the maximum allowable interruption) and RPO (maximum period during which one can lose recording data). However, in the case of an infrastructure that is modest in scope, with a slightly higher budget, it is possible - and often desirable - to build an infrastructure which provides high availability that includes two active servers. The advantages of such a system are twofold: the second server can absorb part of the workload and in case of failure of one of the two machines, the application remains functional without any interruption to service. The VPS that was previously used for hosting the ERP can be repurposed to become a machine for monitoring.

1 – Load Distribution

2 – A distributed infrastructure on two active servers, housed in 2 remote datacenters

3 – Replication of databases between the two servers

4 - Backup to So you Start server

5 - Monitoring

Go Further

If the number of ERP users would grow exponentially, the infrastructure could evolve by separating the database from the application on the front-end servers. Then one could increase the number of database servers, in a manner to distribute the load over a larger number of servers. Moreover, the cost of backups can also be reduced by using RunAbove object storage. The goal, to only pay for storage that corresponds to the amount of data without having to provision hard drive resources and pay for a server whose CPU will be used very little. There is an added security with this solution, by default data is replicated 3 times on RunAbove Storage.

Elements of the infrastructure

Frontend servers

2 Infrastructure range servers
Model EG-32


Order now

Private interconnections between servers

Services included with OVH Infrastructure servers

Backup Servers

Server So you Start Range DISK + Servers
Model BK-8T


Order now


1 VPS Cloud 1


Order now