Our client needed more clarity on the current structure of cloud infrastructure spending. The current amount of spendings has grown faster than their revenue.
Furthermore, the client was also interested in finding cost saving opportunities and be able to correctly anticipate their cost growth as their user base increases.
The client is using the following solutions on AWS
- EC2 for application servers
- S3 for file storage
- ELB for auto-scaling
Data Kernel used a three-pronged approach to helping the client.
First, a linear model was created to estimate the impact of platform scale on infrastructure cost. We combined client data from multiple logging and time-series databases.
The client also received a visualization of their spending structure to see cost hot spots. This immediately allowed them to see which parts of the software needed to be re-architected or optimized.
After being able to assess the current landscape, we turned to simulating what-if scenarios. We created a Monte Carlo simulation tool that enabled our client to assess the impact of tuning infrastructure parameters. We used real-world traffic data to see what the impact of architecture optimizations on scalability is.
The client was able to reduce their infrastructure cost by 10 %. Having comprehensive planning tools allowed them to better grow their platform while minimizing risk.
The simulation allowed them to identify several performance hot spots that they were able to optimize. Thus, they ensured that their platform will have better growth characteristics in the future.
Overall, the client now is able to advance their technology platform with more self-confidence.