Mathematics in the Business: Monte Carlo Simulation.
- ukrsedo
- Jan 12
- 3 min read
Updated: Mar 9
Choosing the right cloud storage provider can be difficult. IT procurement teams often feel overwhelmed by tiered pricing, data transfer fees, and different storage needs.
This blog will study how Monte Carlo simulation can streamline decision-making. By utilizing public pricing from AWS, Azure, and Google Cloud, we’ll demonstrate how simulation aids in forecasting costs, evaluating risks, and determining the best value for your company.
What is Monte Carlo Simulation?
Monte Carlo simulation is a mathematical method for predicting possible outcomes in uncertain situations. It runs a process thousands of times with various random inputs, generating a spectrum of potential consequences that can be examined to grasp probabilities, trends, and risks.
Cloud pricing involves several unknowns:
• Fluctuating storage and data transfer needs.
• Complex pricing structures like tiered storage costs and volume discounts.
• Variable data transfer fees depending on usage.
Monte Carlo simulation helps model these uncertainties, allowing procurement teams to predict costs more accurately.
Choosing a Cloud Provider
Our Needs
• Storage: Monthly usage fluctuates between 50 TB and 100 TB, with an average of 75 TB.
• Data Transfer: Varies from 5 TB to 20 TB per month.
Provider Options
We analyzed three top providers with the following pricing:
Provider | Storage Cost (per GB) | Data Transfer Cost (per GB) | Volume Discount |
AWS | $0.023 | $0.09 | 10% above 90 Tb |
Azure | $0.0184 | $0.087 | 5% above 80 Tb |
Google Cloud | $0.020 | $0.12 | 8% above 85 Tb |
How Monte Carlo Simulation Works
Calculating Costs: For each simulation:
Base storage cost = Storage Usage x Storage Cost.
Data transfer cost = Data Transfer x Transfer Cost.
Volume discounts were applied if storage exceeded the provider’s threshold.
Running Simulations: We asked ChatGPT to run 10,000 iterations to generate a realistic range of costs.
Cost Analysis by Provider
Provider | Median cost | 95th percentile cost | Key observations |
AWS | $9,263.83 | $10,752.97 | Moderate variability, competitive pricing. |
Azure | $8,513.53 | $9,825.29 | Lowest cost with stable pricing. |
Google Cloud | $9,890.24 | $11,164.55 | Higher costs due to transfer fees. |
What is the 95th Percentile?
The 95th percentile represents the cost below 95% of the simulations fall. For example:
• For AWS, 95% of the simulations show costs below $10,752.97.
• This critical metric reflects the upper limit of expected costs in most scenarios.
Monte Carlo Analysis Outcome
Azure offers the lowest median cost, making it an excellent choice for cost-conscious customers.
Why Monte Carlo Simulation is Applicable
Monte Carlo simulation is perfect in this case because it:
Accounts for Uncertainty: Models fluctuating needs and complex pricing.
Highlights Risks: Provides insights into cost variability and worst-case scenarios.
Supports Decisions: Offers data-driven logic rather than relying on assumptions.
Other Considerations Beyond Costs
When selecting a provider, we also have to consider the following:
Performance and Uptime: Does the provider meet our reliability requirements?
Service Integration: Do we need additional AI or machine learning tools?
Vendor Lock-In Risks: How easily can we switch providers if needed?
Ad Astra
Monte Carlo simulation helps procurement manage uncertainty. In our case, Azure was found to be the most cost-effective choice for our needs.
Why not add yet another tool to our arsenal? It's not universal, but it will surely be helpful when the situation presents itself.
Comments