Modernizing database architecture for a custom software development provider’s low-code application.

The Daffodil team implemented several improvements to the existing database architecture of the provider for optimizing the performance of their data management solution. The team leveraged various Amazon RDS solutions to lower database maintenance efforts, reduce infrastructure cost and downtime, and facilitated other capabilities such as the following:

Read-Write Replication

AWS Aurora MySQL was used to enforce a master-slave architecture so that read-write replication could be implemented. With this implementation, high-volume application read traffic could be served while avoiding the need to copy data to the replica nodes. This increased the aggregate read throughput making database scalability a simpler process and drastically reducing the possibility of data loss while scaling.

Auto Scaling 

The data management application’s run on the database was monitored so that the capacity could be automatically adjusted to address scalability requirements. This helped maintain a steady, measurable, and predictable high-availability at the lowest cost possible. As the capacity scaling continued, the associated costs kept decreasing powered by Amazon Storage Auto Scaling.

Security Checks and Remediation

The AWS Security Hub and GuardDuty implementation services were leveraged to perform checks for best security practices. Whenever non-compliant and unencrypted resources as defined by AWS Config rules were found, swift remediation was carried out. This helped better prepare the database for future security threats and bugs through aggregated alerts.

Cost Management

The team utilized the AWS Cost Explorer to get a clear picture of the AWS costs and usage. The costs and usage were visualized at the high level as well as the level of EC2 instances and other low level resources. The top accounts and services and idle instances where the costs were the highest were identified so that proper resource optimization could be implemented and overall costs could be managed.

Load Balancing

ProxySQL was used to create a single endpoint for MySQL connections to utilize the connection pool effectively. It was set up for load balancing for improved availability with automatic failover. In case a hurdle was faced in establishing a connection to one of the database servers or if the server had been shut down for servicing, ProxySQL distributed traffic over multiple servers for quickly switching to a standby database server.