March 7, 2018

Amazon RDS Cost Optimization

Amazon RDS Cost Optimization

The Amazon Relational Database Service (RDS) enables you to have your database inside the AWS cloud without incurring the expenses normally associated with an on-premises infrastructure. This could consequence in noteworthy savings, but Amazon RDS will most likely be the major expense on your AWS bill.

The pricing model isn’t just important for financial reasons, it’s also important when it comes to performance. Failure to understand the pricing model can result in unnecessary expenses. RDS database instance expenses are based on several factors — instance class and storage being the most significant ones.

Database instance cost is the biggest expense for an Amazon RDS instance, but it’s not the only one. There are several factors influencing the total price of Amazon RDS and they need to be included in any cost estimate calculation. The cost of the database instance is based on the instance class and the running time billed by the instance hour. The storage price is the cost of the storage space reserved for the database instance, billed per GB per month.

Your bill will be prorated if you scale your provisioned storage capacity within a month. Provisioned IOPS (PIOPS) storage reserves resources so you are charged regardless of whether or not you used those reserved resources. When you’re using PROVISIONED IOPS storage, you won’t be charged for the throughput and provisioned IOPS rates even though you’re using it.

If you’re not using PROVISIONED IOPS storage, the number of I/O requests per month is included as a cost for Amazon RDS. Write I/O will be doubled if you’re using Multi-AZ deployment for your database instance as Amazon RDS synchronously replicates your data to the standby database instance.

Backup storage is storage associated with a database instance that is used for automatic backups and manual snapshots. Amazon RDS doesn’t charge for backup storage space for up to 100% of your provisioned database storage.

Data transfer, in and out of the Internet, is another expense. You need to factor in the transfer between AWS regions when you’re estimating cost for your database instance; when cross-region replication is implemented, the cost of data transfer between AWS regions will increase.

Amazon RDS Cost Optimization
From a comparison of the price of EC2 instances to the price of equivalent Amazon RDS instances, it’s easy to see that the cost of EC2 is lower. Given that we know instance expenses are the single biggest cost a self-hosted database on an EC2 instance instead of an Amazon RDS database instance.

So using a No-SQL DB Like(Mongodb, Casandra) on EC2 instance with HA and Scalability for Analytics, Repetitive read information and redirecting all the read in the application to fetch the data from No-SQL DB will enable us to reduce the instance type for a large machine to an optimal configuration.

A Simple Mongodb cluster architecture as defined below can sustain heavy loads than RDS and is also highly available and Thanks to Arbiter mechanism built by Mongodb team which will automatically elects the next available instance a primary and helps to check the health of the db instances.


Why No-SQL?

Better performance
NoSQL databases sometimes show better performance, although it depends on the condition and is disputed.

You can add and remove “columns” without downtime. In most SQL servers, this takes a long time and takes up a load of load.

Application Design
It is looked-for to separate the data storage from the logic. If one link and select things in SQL queries, they are mixing business logic with storage.

View/Post Comment