Unique configuration distributed over two data centres
“The high availability of our new configuration is made possible by the geo-redundant storage of data,” explains Alexander Borry, Head of Operations at Artilium. “We chose to distribute our installation over two data centres in two different locations.” This ensures the telecom company can offer optimal availability and continuous service provision. An even distribution of the load on the network also ensures minimal downtime. “If there’s a problem in the primary data centre we can perform a local failover, and if something goes badly wrong locally we can switch over to the disaster recovery site,” adds Bjorn Claeys, Lead Architect at Artilium.
The unique configuration ensures an even distribution of load on the network and guarantees continuous service provision. “On New Year’s Eve for example the load on the network is much greater than normal, but our system can handle that perfectly.”
Disaster recovery rules out data loss
Pieter Vanhove, Senior Database Consultant at Kohera: “Artilium also needs to be able to guarantee full service in the event of a disaster situation. The design makes this perfectly possible. The AlwaysOn failover clustering means the four servers in the primary data centre are in constant communication with each other. This enables us to provide the high availability locally. We’ve also built in AlwaysOn availability groups between the two locations. Whenever there’s a transaction at SQL Server level, for example a new or deleted record, it’s automatically sent to the second site too.” “What makes the Artilium design so unique is the combination of AlwaysOn failover clustering, which ensures high availability, and AlwaysOn availability groups, which provide the disaster recovery,” says Vanhove.
Data warehouse and reporting with minimal load
“We can make these two disaster recovery servers active in the future too,” adds Kristof Van de Putte, Head of Development at Artilium. “This means that we can take real-time reporting from these two servers without needing to load the production servers.” Alexander Borry: “We deliberately chose a full Microsoft environment. Microsoft SQL Server is a reliable product that we’ve never had any problems with and which easily competes with other products. In that respect the Microsoft technology compares very favourably with other providers’ products.”
Logical choice for Microsoft and Kohera
Choosing Kohera was a very obvious decision. “It’s a logical choice considering the successful collaboration we’ve already had with the integrator in the past. Kohera has the biggest team with vast experience of SQL migrations in Belgium, so it can quickly find out what we need,” says Borry. “At the start of the project we all put our heads together to look for the best solution for the implementation strategy in terms of the Artilium design. Then we worked out the implementation plan,” says Vanhove. “I first automated the installation with Microsoft PowerShell because we had to transfer 18 SQL Servers, and attempting that manually was a hopeless task. Once that was done, we performed various tests to check how the system and availability groups were performing. Then we developed and tested the disaster recovery procedure, which is also fully written in PowerShell. We monitored the first migration together with Artilium, and Artilium itself ensured all the other migrations ran smoothly. The entire project was completed in five months.”
Customer appreciates result
Alexander Borry: “Our customer was delighted with the result. Good preparation ensured the implementation ran in a very structured way and we got the desired result. It’s boosted their confidence in us even more.”
About Artilium
Artilium develops real-time, multi-brand solutions that operators in the telecom sector can use to open up their networks to Mobile Virtual Network Operators. The company was looking for an infrastructure that performed well in a future-proof environment and could grow according to the needs of its client, a large Belgian operator. High availability and extensive reporting possibilities were important aspects in the choice for the new hardware. The new servers and software are distributed over two data centres: four in the primary data centre and another two in the recovery data centre. The setup and architecture of both locations is unique in the way in it ensures high availability: optimal accessibility to the network with minimal downtime. IT partner Kohera has built up a vast amount of experience in migrations with SQL Server, and ensured a seamless implementation of the design developed by Artilium.
