Concepts

Uptime

Guaranteeing the availability of your data and applications is Combell's main concern. We can guarantee this availability in several ways.

High availability

The principle of high availability is easy. Every server is duplicated by a second server. If a problem arises on the first server, the second server automatically takes over the activity, without the availability of data or applications being interrupted. The technique intercepts both technical and logical errors.

With fault tolerance, a mirror of the server is constantly run on a second physical machine, just as with high availability. If a problem arises with the server, here too, the service is not interrupted. However, fault tolerance offers no protection against logical errors.
There is also virtual high availability. In this case, the second instance is run on another physical server and the failover from one server to another takes a little longer, which means there is a short service interruption.

Custom work is the best way to guarantee the constant availability of your company data and applications. According to your needs, Combell develops the most adequate solution, by combining high availability, fault tolerance and virtual high availability.

Load balancing & clustering

By using load balancing technology, Combell distributes the workload of your servers optimally across your server farm. This way, we make sure that your infrastructure is as ready as possible to support your business.

There are different methods: a layer 3 load balancer distributes the workload of the servers at the level of the network, a layer 7 load balancer does that at the level of the application layer, combined with SSL or not.

Clustering is a technique that allows to get more out of the existing infrastructure. With clustering, two servers (or more) form a logical entity. With clustering, two database servers act as a logical entity.

A cluster can be active-active, in which case it is distributed across two nodes. With an active-passive cluster, all resources are located on a single node, with a second node as a solution in case the active node should fail. We provide clustering as a failover solution, in order to guarantee high availability of data and applications. Besides traditional forms of clustering based on two nodes, we have built extensive experience with multi-node clustering, which enables us to manage up to three or four database nodes per cluster.

Service Level Agreements

A Service Level Agreement guarantees the service level that the customer agreed with Combell.

This way, the SLA is also part of the series of measures that you take with Combell to ensure the availability of your data and applications. Two essential elements that form an SLA are the response and the recovery waiting times. Combell expresses the guaranteed response and recovery waiting times in minutes. When an incident occurs, we guarantee you that we will take action within a set timeframe and that we will also come up with a solution within a set timeframe. We guarantee you that your environment will be up and running again within the agreed timeframe - in practice, that is generally a couple of minutes after the incident.

Disaster recovery

If, against all odds, the availability of the service should be interrupted, Combell has the necessary procedures to get your environment up and running again as soon as possible.

The conscious choice to use uniform hardware provides an important benefit. Of course, you have enough choice to configure the solution that will fully meet your wishes and your needs. At the same time, this uniform hardware allows Combell to guarantee the availability of an efficient stock of replacement hardware in the data centre. That significantly shortens the recovery waiting times in case of an interruption. But there is more: by doing so, Combell is very efficient, certainly when you compare its services to the guarantees offered by hardware manufacturers in case of a failure. Combell has extremely strict procedures for disaster recovery, partly thanks to the permanent stock of replacement hardware, and partly thanks to several standby servers that are available in Combell's network. At the same time, we can quickly activate new dedicated servers for our customers. For standard configurations, that can be done within 24 hours.

Multi-data centre

Combell's multi-data centre approach offers you the best availability guarantees. Even if a severe calamity should destroy an entire data centre, your application would remain available.

If you choose an active-active configuration, both data centres are available together. If you choose an active-passive configuration, the second data centre acts as a backup site in case something should go wrong in the first data centre.