Client Profile
The client is a digital solutions provider offering cloud-based platforms for online payments, e-GIRO services and procurement management. The company was founded in 2014 and they are operating in Singapore and Malaysia. The client supports businesses with secure transaction processing and streamlined procurement workflows, to automate ordering, invoicing, and payments for businesses, especially in F&B with a focus on efficiency, financial control and real-time data for better decision-making.
Business Challenge
As the AWS environment scaled, cloud costs began increasing month over month. Client observed unpredictable monthly bills, inflated costs and reduced visibility over cloud cost drivers. They sought LogusIMS to help analyze, optimize and bring the spendings under control.
LogusIMS Solution
LogusIMS team conducted detailed analysis of AWS infra setup, billing pattern and identified key factors for cost inflation which could be brought under control as listed below:
- All AWS instances were running on higher pricing model even though cost reduction options were available.
- Numerous servers were exposed to public address resulting in higher costs
- Unnecessary public routing resulted in extra data transfer costs
- Instances running during non-usage periods leading to additional billing
LogusIMS team implemented cost optimization measures aligned with AWS infra workload usage patterns and AWS best practices to address the inflation factors reducing the cost, as detailed below:
- Reserved instances were adopted for instances of known usage pattern to reduce cost (applicable for both EC2 & RDS)
- Network architecture optimization was carefully planned and executed to reduce billing, like migration of EC2/RDS instances to private subnets, routing external access through ALB/VPN endpoints etc. This also improved cloud security posture
- Data transfer cost optimization was done through elimination of unnecessary outbound traffic, reduced inter-service communication and reduced NAT gateway utilization
- Automated start/stop was implemented, through shell scripts, for chosen instances during non-usage periods to save cost