What is clustering? Clustering is the ability to build in a certain amount of redundancy by the replication of hardware.
Simply put, you can keep your company safe from local outages (your server fails, some part of the server fails, your network fails, etc). Now, this can be done with the replication of hardware.
How? We’ll call your main server, “machine A”. Machine A has a lot of internal redundancy built in. It has disk replication where it reads from another disk in the event of a disk failure, it has redundant power supplies each one capable of handling the full load of “machine A”, it has two network ports in the event of failure of one of them. Machine A is a great machine. However, machine A doesn’t have a redundant motherboard where all the components live. The motherboard is called a ‘single point of failure’ in this case. So…. one day, machine A’s motherboard coughs, sputters and dies.
All of the services that machine A provides to your company are gone. These could be mission critical functions like e-mail, database access, directory services, file serving, etc. Your company has stopped moving in some, possibly all, areas.
This could have been avoided with Atlantis Availability. Clustering keeps things running. Clustering allows invisible updating/maintenance of the operating system. (Yes, Atlantis can update your server during your business hours and you, usually, won’t see anything more than a 5 minute hiccup due to switching over to another machine in the cluster.)
Also, there’s different levels of clustering.
Server redundancy – This is where there is a ‘machine B’ where it’s identical to ‘machine A’ and there is software running on both to detect a server failure and move all the services instantly to ‘machine B’. There is only one storage array that is connected to both (or more) servers.
Server and data redundancy – This is where there is a redundant storage array, ‘array B’. Machine A and array A form one entity and machine B and array B form a second entity. The cluster software handles all the writes and replication to both machines. This version is more robust than the first as the same problem exists where the equivalent of a machine motherboard exists in a storage array and can be a ‘single point of failure’
There are local clusters and there are remote clusters that replicate over encrypted communications over distance for even more safety.
And, again, Atlantis is not a hardware vendor. We will make recommendations and we’re not getting kickbacks from vendors. The only thing we will be charging for is the support and services of your machine. This will make your costs lower and we will be your partner.