One of the key innovations introduced by NoSQL databases is a scale-out architecture that lets you expand a database deployment over inexpensive, commodity hardware. Relational databases, on the other hand, forces you to buy a bigger, more expensive server when you need to scale.
The drastically lower cost accounts for NoSQL's growing popularity as a standard in modern application development. For example, deployments with MongoDB, the most popular NoSQL database, typically employ cheap Linux servers that cost $3,000 or less. NoSQL is also particularly well-suited for scaling inexpensively using third-party cloud services. Contrast this with the cost of just one relational server that can range in the six figures.
These cost benefits do come at a trade off in terms of the complexities involved with capacity planning and database provisioning for deployments across multiple servers. MongoDB customers typically want to understand: How many replicas are needed? When should I shard? How much RAM will I need for my working set? Do I use SSD or HDD?
We recommend that you should carefully review the factors that determine capacity requirements such as the volume of queries, data access patterns, indexing, and working set size. We also recommend using Cloud Manager, our easy-to-use management solution that offers automated cloud provisioning. It has the added benefit of integrating with Amazon Web Services so you can automatically provision cloud-hosted machines with an optimal configuration for MongoDB at the time of deployment.
It’s easy to get started and completely risk-free. Cloud Manager is free for the first 30 days. So sign up for Cloud Manager today.