On-demand provisioning provides scalability as needed for cloud-native application components and optimises storage and data management leading to a lowered overall OPEX of 25% and lowered storage costs of 40%

Highlights

• Optimised storage components for the application by maximising for price-performance.
• Implemented autoscaling to maintain performance during traffic spikes.
• Modified the data management architecture to reduce application management overheads.
• Moved the database from block storage to RDS, a database-as-a-service to eliminate DB management tasks.

Overview

The client is a reputed EdTech company offering education management applications and services to higher education institutions and state governments also, across the country. They offer website and software development services, in addition to training and consulting services to their clients.  

Applications comprising over 30+ modules help students and faculty thereby offering complete lifecycle management automation. The application is based on AWS and needs to handle high traffic volumes in a cost-effective manner.  

Requirements

The client’s AWS-based application is accessed by lakhs of users on a daily basis. With the existing architecture, users were experiencing high response times. Because the infrastructure didn’t use autoscaling and on-demand provisioning, the cost of running the application was rising due to a fixed resource allocation which had excess capacity at times, leading to mismanagement and wastage of cloud resources. The client did not have AWS-certified engineers on their team, which made it difficult to manage, provision, and run the operations in a cost-effective manner.

Key objectives:

1. Find the latest generation cloud-native storage alternative to Amazon EBS gp2 SSDs (used for primary storage) which enables faster response times.

2. Implement autoscaling so that the application can handle spikes in traffic while maintaining steady and predictable performance.  

3. Move the MySQL database to an AWS managed service which eliminates the need for maintaining the underlying infrastructure or platform.

This required experienced AWS architects who were up to date with the latest AWS services and could spot optimisation opportunities in the application.

Solution

The client chose Velocis as its cloud service provider to optimise the cloud application. Velocis deployed a team of AWS-certified architects and developers to study and optimise the performance providing, cost-effective storage services. Velocis also proposed to move the MySQL database from EBS to Amazon RDS, which would eliminate database administration overheads

Here are the key factors that enabled optimizing the client’s application:

1. The application servers were using Amazon EBS gp2 SSDs for primary storage. We switched these with the new-generation gp3 SSDs which offered better price-performance. This lowered the application’s response times, and the AWS KMS encryption enabled improved data security. Also configured Amazon Data Lifecycle Manager to automate the creation and retention of Amazon EBS snapshots.

2. Implemented autoscaling for storage and compute components, and configured autoscaling limits in line with the application’s traffic load. Following this, the application can scale automatically during traffic spikes to maintain optimal performance.

3. Moved the MySQL database which was running on EC2 instances to Amazon RDS. This eliminated administration tasks like the need to apply patches and updates, monitoring and scaling the DB manually, and backup and replication.

4. Used Amazon S3 buckets for storing static data and objects, backups, and log files, and S3 Server Access Logging to enable auditability.  

Business Outcomes

With the new strategy, detailed above, Velocis was able to optimise the client’s cloud-native app and align it with AWS best practices. Here are the key benefits that were realized as a result of this engagement:

1. Reduced the cost of storage by 40%, and lowered the time spent on updating and managing production servers by 80%.

2. Reduced the AWS bills associated with running the application by 25% by implementing autoscaling and moving the DB to Amazon RDS.

3. Eliminated the need to manually scale and manage the application.

4. Ensured that the application maintained performance during traffic spikes while utilising the storage and compute resources efficiently.

‍