Basic redundancy schemes for data center engineering systems

[jameswilliams]

In addition to the correct organization of the data center, it is necessary to organize the redundancy of all systems to improve reliability. All redundancy schemes are denoted by the symbol N, which comes from the word "need".

Any system consists of a different number of critical components, each of which is necessary for its functioning. Consider the main redundancy schemes:

Of N . There is no additional redundancy in this option. The reliability of the system directly depends on the quality of the implementation of each structural element. Accordingly, if one component stops working, then the entire system stops functioning. This scheme is outdated and is not used today.
N + 1 . It implies the presence of one spare element that is not used until a problem occurs in the system. As soon as one of the critical components fails, the reserve one will replace it, and this does not affect the operation of the system in any way. The only drawback of the scheme: to carry out repair work, the system will have to be completely turned off.
2N . Its use implies the presence of two more parallel operating elements for each system. In this case, the load will be evenly distributed between the components, and no element will be overloaded. With this option, even if one element of the system completely fails, the other will still work.
2N + 1 . The circuit is identical to 2N, but it is built using one more additional element. This option makes it possible to carry out repair work without stopping the system.
2 (N + 1). The difference from the previous scheme is that even an additional element is duplicated here.
3 / 2N . An advantageous redundancy scheme, since it is loaded by 2/3, and not by ½, which is quite effective. The performance of the system increases and its depreciation decreases.

For more information about: cisco data center