Clustering Types

There are several types of clustering available for different purposes. When you are considering clustering as an option for any purpose below, it is extremely important to remember to select solution which have commecially proven local support and backed life-cycle and is widely used for professional purposes.

It also has to be scalable not only within the cluster, but also able to deploy similar kind of clusters without making every cluster as a own card house project. Cluster solution should always tolerate updates in OS, h/w or in applications without any conflict situations. Clustering is for business critical needs which should be up&running without continuous presence of IT personnel. We highly recommend to choose only commercially proven cross-certified solution to meet and exceed the set requirements.

Load Balancing

Incoming traffic is load-balanced between a set of standard servers. This can be considered more of a farm than a cluster since the server nodes don't typically work together. Load balancing is used for example for high traffic web services. Nordicmind provides Load Balancing cluster as part of its expert services. Please contact Nordicmind for details.

High Performance (HPC)

There are two fundamental clustering types for High Performance Computing.

1. Clustering solution is constructed to run parallel programs where nodes work together and communicate to accomplish their task.

2. Clustering solution is constructed to allow a user on one node to spread his workload transparently to other nodes in the cluster. This is useful for compute intensive, long running jobs that aren't massively parallel.

High Availability (HA)

There are different levels of high availability solutions available. Term "High Availability" is generally used to cover all of those levels and it is sometimes used in context of disaster recovery and data replication as well. However, with advanced HA clustering solutions such as SteelEye LifeKeeper you are able to comply a real high availability cluster with data replication over LAN or WAN.

High Availability clusters are also often known as failover clusters. Resources, most importantly applications and nodes, are monitored and scripts are run when a failure is detected. Scripts are used to fail over IP addresses, disks and filesystems as well as restarting applications. Certified application level support is required from HA clustering software in order to monitor the application health.

Nordicmind Ltd is the value added distributor of SteelEye LifeKeeper, which is one of the leading robust fault resilient clustering solutions in the world. There are also very advanced and robust data replication and disaster recovery solutions available when using SteelEye LifeKeeper.