How to Achieve the Trade-off Between Cost and Performance (with Multi-Objective Analysis)

The Good Spending · 2 years ago · 164 views

In today’s cloud-driven environment, businesses face increasing pressure to balance performance, reliability, and cost in their cloud infrastructure.  This post will explore an example of cloud storage optimization in which a company must meet storage capacity and service level agreement (SLA) requirements while minimizing costs.  We will walk through the process step by step, including how to derive constraints and how Pareto front analysis can help visualize trade-offs between price and performance. Cloud storage optimization problem A company must combine public and private cloud storage options to meet its storage requirements, minimize costs, and ensure an SLA of at least 98%. • Public cloud: $1000 per TB/month, SLA = 95%. • Private cloud: $1400 per TB/month, SLA = 99.8%. • Total storage required: at least 700 TB. • Objective: minimize costs while achieving an overall SLA of 98% or higher. Mathematical problem formulation To solve this optimization problem, we need to define critical variables and constraints. Variables: x  represents the amount of public cloud storage (in TB) y  represents the amount of private cloud storage (in TB). Constraints: The combined storage must be at least 700 TB. x + y>= 700 SLA Constraint: The weighted average SLA from both storage options must meet or exceed 98%. This constraint is derived from the weighted average of the two SLAs: (95x + 99.8y)/(x + y) >= 98 After simplifying, this becomes: y >= 1.67x. This means private cloud storage must be at least 1.67 times that of public cloud storage to meet the SLA requirement. Objective: We want to minimize the total cost: Cost = 1000x + 1400y. Multi-Objective Optimization and Pareto Front. In this case, we primarily focused on minimizing costs, but it also presents a multi-objective optimization scenario where you can balance cost against performance (SLA). Although we constrained the SLA to be at least 98%, businesses may want to explore the trade-offs between improving the SLA and the associated costs. This is where the Pareto front comes into play. The Pareto front represents the set of optimal solutions where no single objective (cost or SLA) can be improved without worsening the other. Moving along the Pareto front highlights the trade-off between achieving a higher SLA and increasing costs. The key points on the Pareto front can show, for example: • $875,131 per month at 98% SLA. • $900,000 per month at 98.4% SLA. • $980,000 per month at 99.8% SLA. By visualizing the Pareto front, decision-makers can decide how much they will invest to achieve a higher SLA. Cost and SLA-optimal solution We will find the optimal solution on the tip of the feasible region formed by SLA and storage constraints: • public cloud storage: 262.5 TB • private cloud storage: 437.5 TB • total cost: $875,131 per month • overall SLA: 98%. This combination satisfies both the storage and SLA requirements while minimizing costs. Conclusion: Balancing cost and performance This case study shows the way to achieve a trade-off between cost and performance.  Through a simple optimization exercise—which looks much scarier than it is—you can find the best balance between minimizing costs and maintaining a high SLA.  Understanding these trade-offs and visualizing them through a Pareto front may help justify the preferred option to decision-makers. To keep receiving new insights and research, please subscribe here.  My Udemy course "Value-based Procurement." My Udemy course "Procurement Lab" My Udemy course "Foundations of Contracts and Outsourcing." My Udemy course "Adaptive Sourcing: Agile Procurement in Practice." More information on this and other exciting topics can be found in "The Technology Procurement Handbook." It represents 23 years of experience, billions of dollars worth of successful sourcing projects, and 1000s of hours spent on research, analysis, and content creation for the most demanding professional readers.   Buy from Kogan Page Buy from Amazon US Buy from Amazon UK Buy from Amazon CA Buy from Amazon AE