Engineers from La Salle-URL share the latest news and projects in the field of network solutions in telematic engineering.

18 March 2019 | Posted by marc.salinero

Clustering in Data Centers

In today's post, it will be explained the "Clustering" topic, that together with virtualization is revolutionizing the Data Centers and completely changing the way they are designed.

Let's start by explaining what the term "Clustering" means. In the world of computer science and networks and Internet, we refer to "Clustering" as a set of physical devices, which connected with a high-speed network act as one logical device. Combining these devices what is achieved are lower costs, facilitate maintenance work and minimize the effects of failure devices.

The main objective of "Clustering" is to be able to achieve high performance and a good load-balance in Data Centers. Thank that this technique allows you to add or remove Data Center devices, it allows you to change the place of devices or replace them to adapt the "hardware" depending on network traffic and servers, so that you can obtain a constant high-performance and a good load-balance.

This technique is very scalable since it can be performed with few or a lot of devices. For example, "Clustering" with few servers can be done in order to run a service while maintaining good machine performance, or it can be done with entire Data Centers in order to run the same services on both Data Centers to provide redundancy and resilience in case of device failures or errors in the network. This last example may seem nonsense, but companies are starting to implement it, combining Data Centers with "Clustering" to form a kind of virtual Data Center on the network. With this companies gain in security in front of failures or technical problems. This is achieved by replicating the contents of a Data Center to another, like a copy of the other, as can be seen in the following figure.

Clustering scheme with Data Centers

In this image you can see the scheme of the Clustering operation in Data Centers.

Applying this technique to combine two Data Centers allows users that are trying to access the services they contain, to be able to access them through either Data Center, achieving a balanced traffic in both. As mentioned above, this method offers a lot of redundancy since it reduces the failure points a lot, since in case of a device crash, it can be replaced with total transparency without affecting the operation of the entire set of devices. In the same way that when you want to change or expand the devices, it is always done with total transparency. And in the case of changes to the services or "software" that are running on certain devices, they would be automatically replicated to other devices in order to maintain an exact replica in all of them to avoid problems in case of failures.

Finally, clarify that virtualization and "clustering" is not the same, since both terms have different goals. But, although their objectives are different, this does not mean that they are incompatible, nowadays ther are techniques that are being developed in order to be able to combine these two technologies and thus be able to take advantage of the advantages that they can offer us separately, to make our Data Centers more powerful and better possible.

Share