Challenge: The telco was looking to optimize its public cloud expenditures on AWS and GCP. The challenge was to identify usage patterns and optimize spending without compromising performance or scalability.
Solution: An ML-based recommendation system was developed to analyze usage patterns and optimize cloud spending. Using a tool called CloudBolt, the system provided baseline ML forecasts similar to on-prem demand prediction to refine and optimize public cloud spend. This approach typically results in around 20% savings on cloud costs in line with industry averages.
Implementation and Results:
Usage Pattern Recognition: The deployed system learned the "normal" usage and behavior patterns of the telco’s cloud resources. By recognizing these patterns, the system could provide demand suggestions for right-sizing the environment.
Cost Savings: Over the course of a year, the telco saved 15% on its AWS and GCP cloud migration costs. This was achieved through automated elasticity, reduction of wastage, and reclamation of temporary resources.
Enhanced Efficiency: The system's automated recommendations and adjustments ensured that the cloud environment was always optimally sized, reducing unnecessary expenditures and improving overall efficiency.
Impact: Implemented in 2021/2022, the ML-based cloud cost optimization system delivered significant cost savings and operational efficiency.
