Getting the most out of AWS for an MVP Solution

Optimisation of how you design and deploy your cloud architecture is important for performance efficiency.

About Our Customer

MyGlobalHome is focused on providing AI-powered Building Intelligence that helps people and organisations invest in, design, construct and occupy green buildings that are environmentally responsible, healthy, and profitable.

MyGlobalHome

Their Use Case

We met MyGlobalHome (MGH) in 2022 through our trusted partner network. MGH had secured investment funding from Innovate UK to build out and develop their Minimum Viable Product (MVP). However, the cloud infrastructure (AWS) design that hosts their solution was not meeting their expectations in terms of efficiency. This was leading to solution performance issues as well as cost inefficiencies. MGH had a deadline to demonstrate progress to their investors in Q3 2022 and needed a skilled team to quickly resolve these efficiency concerns.

Our Solution

Our team has extensive experience working and designing for different cloud infrastructures including Amazon Web Services (AWS).

For MGH's cloud infrastructure optimisation, we deployed a couple of cloud architects to conduct the assessment. We began with a discovery phase to assess the current state of the infrastructure, determine the necessary optimisations, and make the necessary infrastructure adjustments.

Our solution encompassed several key improvements:

  1. Comprehensive Cost Breakdown:

    • Created a detailed cost analysis that categorized expenses into fixed costs, gradually increasing costs, and those scaling linearly with customer growth.

  2. Optimised Log Management:

    • Reduced log capture levels in CloudWatch.

    • Collaborated with applications to eliminate unnecessary or improperly levelled log entries.

  3. Enhanced AWS IOT Infrastructure:

    • Partnered with the AWS team to identify and resolve bugs in the Greengrass client, which could lead to unnecessary high usage of AWS IOT resources.

  4. Efficient API Call Management:

    • Batched updates to AWS IOT Shadows, reducing the frequency of updates to align with application requirements, thereby saving on API call costs.

  5. Selective Metric Logging:

    • Limited metric logging to a subset of devices, ensuring we obtained necessary data from a representative sample without overwhelming the system.

  6. Optimised Web Hosting:

    • Transitioned static web applications to be hosted via CDN instead of active compute services.

  7. Cost Reduction in DynamoDB:

    • Minimised payload sizes to reduce DynamoDB costs.

    • Worked with applications to decrease unnecessary DynamoDB calls.

  8. Optimised Data Ingestion:

    • Fine-tuned Kinesis data ingestion pipelines to handle traffic efficiently.

The Results

The AWS infrastructure optimisations contributed to making MGH’s demonstration to their investors a great success.

Overall, we identified a potential savings of approximately 40% per home with a second round of optimisations.

The optimisations we implemented have already reduced costs from $938 in December 2022 to the current level of $571, despite a 25% increase in the customer base.