Server Cluster

DomainGurus knows that there is actually no easy way to explain what a server cluster is, unfortunately we have to use technical terminology to refer to them constantly.

That is why before you make the wrong purchase we take the time to directly attend your particular case, we read your needs posted to our emails in order to build a quote and to answer the questions you might have about the subject.

So we will try to explain in easy words what a cluster is, we will cut the jargon and go straight to the point.

One could imagine a cluster as a team of servers or a team of computers commonly called nodes, in other words what we are we try to achieve is to make a whole bunch of computers work as one.

You can think of it as a football team, although their more than one player per team, and they would like to play their way they are a team and they are trained to work as a team, in other words as one, and computers aren't any different.

There are 3 principle task you can achieve by implementing a server cluster:

  • Availability
  • Load Balancing
  • High Performance

Availability, if we follow the football team example, availability would be more like having players on the bench field, if one of the players gets hurt another one will take their place, but unlike football computers have a little twist there is no rule that tells you how many computers your cluster can have, or if all of them can o cant be on the playing field, what we mean by this is that with computers you would have all of your players (servers) playing (running) at the same time and if one of them gets compromise in any way the rest will take over and keep working so that no one besides the coach (node manager) is able to tell who is missing or why is that whom missing.

Load Balancing, now, imagine this tall gigantic guy of 500 pounds of weight (giant task) in order to take him down you would need more than one single player (server) thats why you should have your whole team pushing him down (cluster), there is no better example than this one, because that is really how it works on the real life in computing.

High Performance, Now lets say its the end of the game, you only have 5 seconds to score, and time is running, in order for one of your players (servers) to score (finish a task) you would have to rely on the whole team (cluster) and the reason for that being is the fact that one player (server) is not going to do everything by himself, we could even say it would be impossible.

If we can explain Linux Server Clustering as it if it where football or any other sport or game, then just imagine how easy we could make things for you when telling us about that development you are working on and you need a to handle a whole bunch of tasks, or you just need to know where you're at but also need to understand what we are talking about, is just matter of cutting the jargon as we just did and explaining in a different way.

The cluster idea is not particular from computing but also applicable to many facts of life.


In a perfect cluster, clients don't know they are working on cluster, applications don't know they are working on a cluster, not even cluster nodes know they are working on a cluster only the node manager who has to be cluster aware in order to
tell when we bump into a cluster node death, other than that your applications can be crystal clear to you and your customers and more reliable.