S3 Optimization Glossary

Understanding these key terms is essential for effective storage cost management and observability. Each term represents a core concept that s3lim is designed to detect and measure.

Small File Trap

An architectural pitfall where an application stores millions of small objects (typically <128KB). This results in PUT/GET request costs and S3 minimum storage charges overpowering the actual storage cost of the data.

Ghost Versions

Non-current object versions and expired delete markers that occupy storage space but are not visible during standard S3 listing operations. These can account for a significant portion of an older bucket's bill.

Lifecycle Drift

A state where objects have remained in expensive storage tiers (like Standard) longer than their access pattern warrants, failing to transition to IA or Glacier as defined by organizational policy.

Zero-Egress Analysis

A security pattern where data is analyzed entirely within the AWS account or VPC where it resides. No data is transferred across regions or to external third-party SaaS providers, eliminating data transfer costs and security risks.

Top-K Aggregation

A probabilistic algorithm used by s3lim to identify the most significant 'heavy hitter' prefixes or objects within a massive dataset using constant memory, regardless of the number of unique keys.