Why use shards with MongoDB?

Question

We started using MongoDB at work. So far, so good, however, I was asked to evaluate whether MongoDB can perform copying and smoothing at the same time. After looking for some time, I found that yes, but this can lead to the fact that fragments should be avoided if it is really necessary.

I read somewhere that with MongoDB you really need to have a lot of document (for example, billions) in order to start thinking about Sharding, however I do not understand this, since the organization and size of the average document are not mentioned in these "statistics".

The criteria that we have now is that the RAM should be more than the memory load needed to complete the request.

This leads me to this question: What are the instructions for MongoDB and the SQL server to find out what RAM is used for the query?

What are the basic requirements for using Sharding with MongoDB?

Can someone point me to an article, a post, or something when a person explains why they used Sharding with MongoDB?

How does MongoDB handle Auto-Sharding?

Thanks.

Answer 1

From what I understand, your working set (the data that you most often access) should correspond to RAM for good performance. If this is not the case, then selection may be an option.

Replica sets allow you to distribute readings across multiple sets, while a shape can also serve to distribute records.

O'Reilly has a book called “scaling mongodb,” which explains how to use replica sets and fragments.

You can also see this page: http://www.mongodb.org/display/DOCS/Sharding+Introduction