Cloud Pricing: On-Demand, Reserved, and Spot Instances
Cloud pricing is like booking a flight. Pay full price for flexibility (On-Demand), get a discount for committing (Reserved), or bid on empty seats for a huge discount but risk getting bumped (Spot). This choice dictates your infrastructure cost.
WHY IT EXISTS: Cloud providers have massive data centers with fluctuating demand. Pricing models help them sell excess capacity and allow users to optimize costs based on their workload's needs. This strategy lets providers monetize idle hardware while giving customers financial incentives to commit to long-term usage.
THE MENTAL MODEL: Think of it like airline tickets. On-Demand is a last-minute, full-price ticket with maximum flexibility. Reserved Instances are like buying a non-refundable ticket months in advance for a steep discount. Spot Instances are like flying standby; you get an incredible price but risk getting bumped if a full-fare passenger needs your seat.
HOW IT WORKS: The three main models are On-Demand, Reserved, and Spot. First, On-Demand lets you pay by the hour or second with no commitment, offering the most flexibility at the highest price. Second, Reserved Instances (RIs) or Savings Plans give you a large discount (up to 75%) in exchange for a 1 or 3-year commitment to a certain level of usage. Third, Spot Instances let you bid on spare compute capacity for up to 90% off, but the provider can reclaim them with only a short warning, typically two minutes.
WHEN TO USE IT: Use On-Demand for development, testing, and unpredictable, spiky workloads. Use Reserved Instances for your stable, predictable production workloads like databases and core application servers. Use Spot Instances for fault-tolerant, stateless tasks like batch processing, data analysis, or CI/CD jobs that can handle interruptions.
WHEN NOT TO USE IT: The biggest mistake is using the wrong model for your workload. Don't run a 24/7 database on On-Demand; you're burning money. Don't buy a 3-year Reserved Instance for a temporary project. And never run a critical, stateful service that cannot tolerate interruption, like a primary database, on a Spot Instance.
ONE CANONICAL EXAMPLE: A media company runs its core website on Reserved Instances to handle baseline traffic cost-effectively. When a news story breaks and traffic surges, it scales out using On-Demand instances. Overnight, it processes large video files using a massive, cheap fleet of Spot Instances, which is tolerant of interruptions.
Read the original → en.wikipedia.org
Get five bites like this every day.
Tezvyn delivers a daily feed of 60-second tech bites with quizzes to lock in what you learn.