AWS Cost Optimisation

What Is AWS Cost Optimisation?

AWS cost optimisation is the systematic process of analysing, controlling, and reducing Amazon Web Services spending – without compromising application performance or availability. For mid-market companies with monthly AWS bills of €5,000 to €100,000, cost optimisation is a continuous lever that typically enables 20–40% savings.

The challenge lies in the complexity of AWS pricing models: over 200 services each with their own billing models, regional price differences, and hidden data transfer costs make systematic optimisation essential.

The Most Common AWS Cost Traps

• Oversized Instances: EC2 instances with significantly more CPU and RAM than needed. Studies show that 30–40% of all cloud instances are oversized.
• Unused Resources: Orphaned EBS volumes, old snapshots, unused Elastic IPs, and detached load balancers generate ongoing costs without benefit.
• Data Transfer Costs: Outbound traffic (egress) is charged per GB on AWS. Cross-region and cross-AZ traffic can cause significant costs.
• Dev/Test Environments 24/7: Development and test environments running around the clock despite being needed only during business hours.
• Missing Tagging: Without consistent tagging, it is impossible to allocate costs to teams, projects, or environments.

Reserved Instances and Savings Plans

Reserved Instances (RIs)

Reserved Instances offer discounts of 30–72% compared to on-demand pricing. You commit for one or three years to a specific instance family, size, and region. RIs are ideal for baseline workloads that run continuously – such as database servers or core applications.

Practical example: A mid-market online retailer runs three m6i.xlarge instances for their web shop. On-demand costs approximately $1,100/month. With 1-year RIs (All Upfront), costs drop to approximately $700/month – a 36% saving.

Savings Plans

Savings Plans are more flexible than RIs: you commit to a specific spend per hour (e.g., $10/hour) rather than a specific instance. Savings Plans apply across EC2, Fargate, and Lambda. Discounts are similar to RIs (up to 72%), but you retain the flexibility to change instance types and sizes.

AWS Cost Explorer and Budgets

AWS provides powerful native cost analysis tools:

• AWS Cost Explorer: Visualises your costs and usage over time. Provides recommendations for right-sizing and Reserved Instances. Enables analysis by service, region, tag, and account.
• AWS Budgets: Define monthly or annual budgets and receive alerts before exceeding them. You can configure budget alerts for individual services, tags, or accounts.
• AWS Cost Anomaly Detection: Automatically detects unusual cost increases and notifies you via email or SNS.

Tagging Strategy for the Mittelstand

A consistent tagging strategy is the foundation of any cost optimisation. At minimum, every resource should have these tags:

• Team/Department: Who is responsible? (e.g., team:ecommerce)
• Environment: Dev, staging, or production? (e.g., env:prod)
• Project: Which project does the resource belong to? (e.g., project:webshop)
• Cost Centre: For internal allocation. (e.g., costcenter:CC-4711)

Enforce tagging via AWS Organizations Service Control Policies (SCPs) or AWS Config Rules to ensure no untagged resource is created.

Practical Optimisation Measures

• Right-Sizing: Use AWS Compute Optimizer to identify underutilised instances and switch to appropriate sizes.
• Spot Instances: For fault-tolerant workloads like batch processing, CI/CD runners, or data analytics. Savings of up to 90%.
• Scheduling: Automate starting and stopping dev/test environments with AWS Instance Scheduler or Lambda functions.
• Storage Tiering: Move infrequently accessed data from S3 Standard to S3 Infrequent Access or Glacier. Intelligent Tiering automates this.
• Graviton Instances: AWS Graviton (ARM-based) offers up to 40% better price-performance than comparable x86 instances.